jackkuo commited on
Commit
f85675b
·
verified ·
1 Parent(s): 09082e5

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. .gitattributes +1 -0
  2. 1NE0T4oBgHgl3EQf_gL8/content/tmp_files/2301.02829v1.pdf.txt +284 -0
  3. 1NE0T4oBgHgl3EQf_gL8/content/tmp_files/load_file.txt +155 -0
  4. 1tAyT4oBgHgl3EQfPfba/content/tmp_files/2301.00027v1.pdf.txt +0 -0
  5. 1tAyT4oBgHgl3EQfPfba/content/tmp_files/load_file.txt +0 -0
  6. 2tAyT4oBgHgl3EQfPvai/content/tmp_files/2301.00031v1.pdf.txt +781 -0
  7. 2tAyT4oBgHgl3EQfPvai/content/tmp_files/load_file.txt +407 -0
  8. 3dE3T4oBgHgl3EQfPwmd/content/tmp_files/2301.04406v1.pdf.txt +794 -0
  9. 3dE3T4oBgHgl3EQfPwmd/content/tmp_files/load_file.txt +0 -0
  10. 49AyT4oBgHgl3EQfcPf_/content/tmp_files/2301.00281v1.pdf.txt +305 -0
  11. 49AyT4oBgHgl3EQfcPf_/content/tmp_files/load_file.txt +160 -0
  12. 4NFAT4oBgHgl3EQfExwU/content/tmp_files/2301.08423v1.pdf.txt +0 -0
  13. 4NFAT4oBgHgl3EQfExwU/content/tmp_files/load_file.txt +0 -0
  14. 59E3T4oBgHgl3EQfQwmQ/content/tmp_files/2301.04416v1.pdf.txt +237 -0
  15. 59E3T4oBgHgl3EQfQwmQ/content/tmp_files/load_file.txt +236 -0
  16. 8dFLT4oBgHgl3EQftC_m/content/tmp_files/2301.12150v1.pdf.txt +540 -0
  17. 8dFLT4oBgHgl3EQftC_m/content/tmp_files/load_file.txt +0 -0
  18. 9dFLT4oBgHgl3EQfuS8F/content/tmp_files/2301.12154v1.pdf.txt +1585 -0
  19. 9dFLT4oBgHgl3EQfuS8F/content/tmp_files/load_file.txt +0 -0
  20. 9dFPT4oBgHgl3EQfYjT_/content/tmp_files/2301.13074v1.pdf.txt +2263 -0
  21. 9dFPT4oBgHgl3EQfYjT_/content/tmp_files/load_file.txt +0 -0
  22. 9tE4T4oBgHgl3EQfDQub/content/tmp_files/2301.04868v1.pdf.txt +2267 -0
  23. 9tE4T4oBgHgl3EQfDQub/content/tmp_files/load_file.txt +0 -0
  24. ANFKT4oBgHgl3EQfVi5k/content/tmp_files/2301.11788v1.pdf.txt +679 -0
  25. ANFKT4oBgHgl3EQfVi5k/content/tmp_files/load_file.txt +0 -0
  26. AtAzT4oBgHgl3EQfF_uW/content/tmp_files/2301.01021v1.pdf.txt +1024 -0
  27. AtAzT4oBgHgl3EQfF_uW/content/tmp_files/load_file.txt +0 -0
  28. AtFAT4oBgHgl3EQfrx4U/content/tmp_files/2301.08654v1.pdf.txt +977 -0
  29. AtFAT4oBgHgl3EQfrx4U/content/tmp_files/load_file.txt +0 -0
  30. INAyT4oBgHgl3EQfffiq/content/tmp_files/2301.00342v1.pdf.txt +855 -0
  31. INAyT4oBgHgl3EQfffiq/content/tmp_files/load_file.txt +0 -0
  32. LNE4T4oBgHgl3EQf8A7c/content/tmp_files/2301.05345v1.pdf.txt +1246 -0
  33. LNE4T4oBgHgl3EQf8A7c/content/tmp_files/load_file.txt +0 -0
  34. LtFRT4oBgHgl3EQf2Di0/content/tmp_files/2301.13659v1.pdf.txt +1263 -0
  35. LtFRT4oBgHgl3EQf2Di0/content/tmp_files/load_file.txt +0 -0
  36. Q9E3T4oBgHgl3EQfyQvd/content/tmp_files/2301.04719v1.pdf.txt +0 -0
  37. Q9E3T4oBgHgl3EQfyQvd/content/tmp_files/load_file.txt +0 -0
  38. RtAyT4oBgHgl3EQfVPc7/content/tmp_files/2301.00139v1.pdf.txt +0 -0
  39. RtAyT4oBgHgl3EQfVPc7/content/tmp_files/load_file.txt +0 -0
  40. WtFPT4oBgHgl3EQfrjVl/content/tmp_files/2301.13145v1.pdf.txt +820 -0
  41. WtFPT4oBgHgl3EQfrjVl/content/tmp_files/load_file.txt +516 -0
  42. Y9E3T4oBgHgl3EQfcgoK/content/tmp_files/2301.04525v1.pdf.txt +1157 -0
  43. Y9E3T4oBgHgl3EQfcgoK/content/tmp_files/load_file.txt +0 -0
  44. Z9E5T4oBgHgl3EQfDg58/content/tmp_files/2301.05406v1.pdf.txt +1689 -0
  45. Z9E5T4oBgHgl3EQfDg58/content/tmp_files/load_file.txt +0 -0
  46. ZNAyT4oBgHgl3EQfifg0/content/tmp_files/2301.00395v1.pdf.txt +2656 -0
  47. ZNAyT4oBgHgl3EQfifg0/content/tmp_files/load_file.txt +0 -0
  48. b9FAT4oBgHgl3EQfXx3V/content/tmp_files/2301.08536v1.pdf.txt +0 -0
  49. b9FAT4oBgHgl3EQfXx3V/content/tmp_files/load_file.txt +0 -0
  50. eNFAT4oBgHgl3EQf7B5d/content/tmp_files/2301.08742v1.pdf.txt +234 -0
.gitattributes CHANGED
@@ -153,3 +153,4 @@ etE2T4oBgHgl3EQfGgZ8/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -tex
153
  4NE1T4oBgHgl3EQfSgOT/content/2301.03067v1.pdf filter=lfs diff=lfs merge=lfs -text
154
  9tAyT4oBgHgl3EQfqPjX/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
155
  n9E5T4oBgHgl3EQfjw-t/content/2301.05658v1.pdf filter=lfs diff=lfs merge=lfs -text
 
 
153
  4NE1T4oBgHgl3EQfSgOT/content/2301.03067v1.pdf filter=lfs diff=lfs merge=lfs -text
154
  9tAyT4oBgHgl3EQfqPjX/vector_store/index.faiss filter=lfs diff=lfs merge=lfs -text
155
  n9E5T4oBgHgl3EQfjw-t/content/2301.05658v1.pdf filter=lfs diff=lfs merge=lfs -text
156
+ o9FMT4oBgHgl3EQf7jFB/content/2301.12464v1.pdf filter=lfs diff=lfs merge=lfs -text
1NE0T4oBgHgl3EQf_gL8/content/tmp_files/2301.02829v1.pdf.txt ADDED
@@ -0,0 +1,284 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Dark Matter and MOND: Two sides of the same coin?
2
+ D. F. Roscoe (The Open University; [email protected])
3
+ ORCID: 0000-0003-3561-7425
4
+ 1
5
+ arXiv:2301.02829v1 [astro-ph.GA] 7 Jan 2023
6
+
7
+ Abstract
8
+ It has recently been reported that the application of convolutional neural-network tech-
9
+ niques to infer the dark-matter distribution in the local cosmos has revealed how it follows
10
+ the D ≈ 2 hierarchical distribution of galaxies in the locality, rather than exhibiting the
11
+ expected homogeneity throughout the IGM. Taken at face value, this implies that the Hub-
12
+ ble Law, observed to be followed on scales which are deep inside the observed hierarchical
13
+ structures, can no longer be assumed to arise from universal expansion. So, if not universal
14
+ expansion, then what?
15
+ As a possibility, it has been recognized for a considerable time that if the lower cut-off
16
+ scales of a D ≈ 2 hierarchical cosmos are identified with the scales of a typical galaxy, then
17
+ gravitational redshift automatically follows the Hubble Law with Hg ≈ 70 km/sec/Mpc.
18
+ Inter alia, this suggests a model of galaxy formation in a D ≈ 2 hierarchical IGM in which
19
+ all of the material M0 within a sphere R0 coalesces about a unique center so that hierachical
20
+ symmetry is broken on the scale (M0, R0).
21
+ Putting these things together leads unambiguously to the conclusion that, in an hierachical
22
+ cosmos, the Dark Matter hypothesis and Milgrom’s MOND hypothesis are two sides of the
23
+ same coin.
24
+ 2
25
+
26
+ 1
27
+ Introduction:
28
+ It is now widely accepted that on scales up to about 200 Mpc galaxies are distributed in a quasi-
29
+ fractal D ≈ 2 fashion. For fairly recent work see Tekhanovich & Baryshev (2016), but many
30
+ others have contributed over the years.
31
+ It then becomes a point of considerable significance that the Hubble Law is well established
32
+ on scales that are deep inside the accepted fractal structure of the general galaxy distribution.
33
+ This very interesting circumstance is the primary evidence supporting the idea that the IGM is
34
+ largely populated by an homogeneous distribution of dark matter on the small scales required
35
+ - for, without homogeneity, the linear nature of Hubble’s Law cannot be understood within the
36
+ context of universal expansion.
37
+ It is for this reason that the paper of Hong et al (2021) caused so much consternation: specifically,
38
+ the authors used state-of-the-art convolutional neural-network techniques combined with modern
39
+ positional and peculiar velocity data to compute and map the local dark matter distribution.
40
+ Against expectation, this distribution is found to trace the hierarchical distribution of galaxies
41
+ very closely - there is no indication of homogeneity, and hence no indication that the Hubble
42
+ Law can be understood in terms of universal expansion.
43
+ The only immediately plausible alternative is some form of gravitational redshift: Baryshev
44
+ et al (1998) point out that inside a D = 2 hierarchical galaxy distribution (with an assumed
45
+ homogeneous distribution of dark matter) the gravitational part of redshift is also purely linear
46
+ with distance and cannot be distinguished from the expansion component. But if the results of
47
+ Hong et al (2021) are to be taken at face value, then any contribution to redshift from expansion
48
+ must manifest itself as a departure from linearity. Since such a departure is not observed then,
49
+ according to the results of Hong et al, there can be no expansion effect at all.
50
+ This line of argument is reinforced by the further observation of Baryshev et al that if the
51
+ lower cut-off mass and length scales of the hierarchy are identified with the mass and length
52
+ scales of the typical galaxy, then a gravitational redshift of Hg ≈ 70 km/sec/Mpc is to be ex-
53
+ pected.
54
+ Inter alia, the foregoing considerations suggest a process of galaxy formation according to which
55
+ an isolated galactic object can be modelled as a finite bounded spherically symmetric peturbation
56
+ of the hierarchical IGM (assumed in the first instance to be a mix of baryonic and non-baryonic
57
+ mass) - this automatically entails that all of the mass M0 within the sphere R0 has coalesced
58
+ around a unique centre so that fractal symmetry is broken on the scales of (M0, R0).
59
+ 2
60
+ Consequences on the lower cut-off scales:
61
+ From these general considerations we may conclude:
62
+ 1. The lower cut-off radial and mass scales (M0, R0) must behave according to
63
+ M0 = 4πR2
64
+ 0ΣF
65
+ (1)
66
+ 3
67
+
68
+ where ΣF is the mass surface density of the D = 2 hierarchical mass distribution in the
69
+ local cosmos;
70
+ 2. Since galaxies in general appear to be stable structures, there must be an equilibrium
71
+ constraint at the lower cut-off scales of the hierarchy. Using simple Newtonian arguments,
72
+ we show in appendix §A that equilibrium at these lower cut-off scales requires:
73
+ V 2
74
+ 0
75
+ R0
76
+ = aF ≡ 4πGΣF
77
+ (2)
78
+ where aF is the characteristic acceleration scale associated with ΣF;
79
+ 3. The relationship
80
+ V 4
81
+ 0 = aF GM0 ,
82
+ (3)
83
+ which is formally identical to the Baryonic Tully-Fisher Relationship (BTFR), is now de-
84
+ rived directly by eliminating R0 between (1) and (2).
85
+ It is to be noted that whilst (3) is formally identical to the BTFR of Milgrom’s MOND, it differs
86
+ fundamentally in the assumption (expressed in the last paragraph of §1) that M0 is an unknown
87
+ mix of baryonic and non-baryonic mass whereas, by definition, the BTFR asserts that this mass
88
+ is purely baryonic.
89
+ 3
90
+ Empirical support for the BTFR hypothesis
91
+ 3.1
92
+ The analysis of Lelli, McGaugh & Schombert (2016B)
93
+ It has only recently been possible to explore the BTFR hypothesis in a statistically rigorous
94
+ fashion. Specifically, the SPARC sample of Lelli, McGaugh & Schombert (2016A) contains high
95
+ quality rotation curves and high quality modern surface photometry at 3.6 µm for a sample of 175
96
+ nearby disk galaxies. The high quality of the surface photometry over this sample allowed Lelli,
97
+ McGaugh & Schombert (2016B) to construct photometric models of baryonic mass distributions
98
+ in that particular subsample of 118 disks which also had rotation curves extending to flatness,
99
+ making it ideal for a statistically rigorous testing of the BTFR hypothesis.
100
+ Subsequently, the authors used regression analysis techniques to demonstrate how the subsample
101
+ really does fit the BTFR with very small scatter. In this way, they argued that the observed
102
+ scatter is sufficiently below the instrinsic-scatter expectations of ΛCDM cosmology to present a
103
+ fundamental difficulty for that cosmology and for the associated idea of dynamically significant
104
+ quantities of non-baryonic matter in the generality of galaxy disks.
105
+ Since (3) is derived from the hypothesis that galaxies form by coalescing in a stable way out
106
+ of the D = 2 hierarchical IGM, this result implies that the IGM itself consists primarily of
107
+ undetected baryonic matter and so, in effect, (3) itself represents a derivation of the hitherto
108
+ empirical BTFR from a fundamental theoretical position.
109
+ 4
110
+
111
+ 3.2
112
+ The estimation of (aF, ΣF)
113
+ The data used by Lelli, McGaugh & Schombert (2016B) is available as an on-line data-sheet giv-
114
+ ing estimates for the photometrically modelled baryonic masses M0 and flat rotation velocities
115
+ V0 for the 118 disk galaxies. Given this data, an alternative demonstration supporting the BTFR
116
+ hypothesis is provided by showing how the hypothesis, applied differently to the data, yields a
117
+ very sharp estimate of the characteristic acceleration parameter aF, thereby demonstrating how,
118
+ for all practical purposes, its value is identical to that of Milgrom’s critical acceleration parame-
119
+ ter, a0.
120
+ In order to estimate aF ≡ 4πGΣF from this data, we rearrange (3) as
121
+ V 4
122
+ 0
123
+ GM0
124
+ = 4πGΣF ≡ constant
125
+ and hence form the empirical sample distribution
126
+ J ≡
127
+ � V 4
128
+ 0i
129
+ GM0i
130
+ , i = 1...118
131
+
132
+ .
133
+ Then, from J, we generate N = 10000 bootstrapped distributions, ˆJi, i = 1..N in the usual way.
134
+ For each ˆJi we then compute its geometric mean, ˆaFi say, to obtain, finally, the distribution
135
+ AF ≡ (log ˆaFi, i = 1..N) .
136
+ The density distribution of AF is given in figure 1 from which it is clear that the estimate for
137
+ aF is very tightly constrained around the modal value of 1.3 × 10−10 mtrs/sec2 which, for all
138
+ practical purposes, is identical to Milgrom’s value of MOND’s critical acceleration parameter
139
+ a0. This estimate of aF corresponds to ΣF ≈ 0.15 kg/mtr2 for the mass surface density of the
140
+ hierarchical cosmos.
141
+ The grey curve in figure 1 arises from the same analysis but applied to shuffled velocity and
142
+ mass data. It is clear that the signal so powerfully present in the unshuffled data is destroyed
143
+ by shuffling. We conclude that, for all practical purposes the signal for the mass surface density
144
+ ΣF ≈ 0.15 kg/mtr2 in the hierarchical cosmos is real.
145
+ 5
146
+
147
+ Dotted = 1.1 × 10−10 mtrs/sec2
148
+ Mode = 1.3 × 10−10 mtrs/sec2
149
+ Dashed = 1.6 × 10−10 mtrs/sec2
150
+ 0
151
+ 5
152
+ 10
153
+ 15
154
+ log(aF)
155
+ Density
156
+ Distribution of bootstrapped geometric means
157
+ Figure 1: Solid black curve = density distribution of log ˆaF. Solid grey curve arises when velocity
158
+ and mass data are shuffled with respect to each other. The signal represented by the black curve
159
+ is destroyed on the shuffled data.
160
+ 4
161
+ Full circle to MOND
162
+ Taking the results of Hong et al (2021) at face value, together with the results of Lelli, McGaugh
163
+ & Schombert (2016B) and observations of Baryshev et al (1998), it has been shown that the Dark
164
+ Matter of modern astrophysics, rather than being an homogeneous distribution of non-baryonic
165
+ matter, can reasonably be identified as a D = 2 hierarchical distribution of undetected baryonic
166
+ matter which, as the BTFR (3) shows, provides exactly the dynamical support for galaxies that
167
+ the original non-baryonic Dark Matter hypothesis was formulated for in the first place.
168
+ 6
169
+
170
+ However, hitherto the prominence of the originally empirically derived BTFR has rested en-
171
+ tirely upon the fact that it is central to the architecture of MOND Milgrom (1983a,b,c). To see
172
+ this, (3) gives directly
173
+ V 2
174
+ 0
175
+ R0
176
+ =
177
+ √aF GM0
178
+ R0
179
+ which, in effect, MOND extrapolates by hypothesis to give g = √aF gN, where gN ≡ GM0/R2, as
180
+ the effective gravitational force for all R ≥ R0 . Consequently, MOND automatically receives the
181
+ direct interpretation as a first-order descriptor of gravitational dynamics in a D = 2 hierarchical
182
+ IGM.
183
+ In short, in an hierachical cosmos, the Dark Matter hypothesis and Milgrom’s MOND hypothesis
184
+ are two sides of the same coin.
185
+ 5
186
+ The nature of baryonic Dark Matter?
187
+ We have argued that the Dark Matter of the IGM is distributed in a D = 2 hierarchy and consists
188
+ of undetected baryonic material. So, the immediate question is: how can such a distribution of
189
+ baryonic material remain undetected? There are three potential strands to the answer, easily
190
+ conceived to be acting in concert.
191
+ 5.1
192
+ Conventional possibilities
193
+ Given that the IGM forms a D = 2 hierarchy, then its volume density in a spherical volume of
194
+ radius R tends to zero as R → ∞, making its detection at large radii intrinsically difficult.
195
+ Furthermore, it can reasonably be assumed that the IGM is at least close to being in thermal
196
+ equilibrium with the general background, again making its detection against the background also
197
+ intrinsically difficult.
198
+ 5.2
199
+ Unconventional possibilities
200
+ Until recently, it has always been assumed that there is no such thing in nature as a perfect, or
201
+ near perfect, blackbody absorber - the reason being that no such thing had ever been observed.
202
+ However, Mizuno et al (2009) showed how to fabricate, from agglomerations of single-walled
203
+ carbon nanotubes (SWCNTs), material distributions having specific bulk statistics which act
204
+ as near-perfect blackbody absorbers (emissivity > 0.98) across a very wide range of incident
205
+ wavelengths from UV at 200nm to the far IR at 200µm.
206
+ This behaviour has been shown to be independent of the specific properties of the individual
207
+ SWCNTs, but is rather a consequence of the bulk statistical characteristics of the fabricated
208
+ SWCNT distributions. We know that many allotropes of carbon exist in interstellar space and
209
+ these must to some extent be blown into the IGM from the generality of galactic interiors. It is a
210
+ short step to visualizing the existence of clouds of SWCNTs dispersed throughout the hierachical
211
+ 7
212
+
213
+ IGM containing sub-populations which, when viewed in projection along any given line of sight,
214
+ possess the bulk statistical characteristics required to mimic the properties of the fabricated
215
+ SWCNT distributions of Mizuno et al (2009).
216
+ In this way, it is possible to conceive how SWCNT clouds within the IGM have the potential to
217
+ act as ‘dispersed near-perfect blackbody objects’ making them virtually undetectable.
218
+ 6
219
+ Summary and conclusions
220
+ There is general agreement that the distribution of galaxies in particular is quasi-fractal D ≈ 2
221
+ out to about 200 Mpc and Baryshev et al (1998) has pointed out that gravitational redshift in
222
+ such an hierarchical cosmos will follow the Hubble Law. Furthermore, these authors point out
223
+ that if the lower cut-off scales of the hierarchy are identified with the mass and radial scales of
224
+ the typical galaxy, then Hg ≈ 70 km/sec/Mpc is to be expected.
225
+ Notwithstanding the hierarchical distribution of galaxies in the local cosmos, the conventional
226
+ view holds that the IGM itself consists of non-baryonic Dark Matter, the assumed homogeneous
227
+ distribution of which makes the Hubble Law a consequence of universal expansion. It is for this
228
+ reason that the paper of Hong et al (2021), which computes and maps the distribution of Dark
229
+ Matter in the local cosmos, caused so much consternation: specifically, they report that the
230
+ distribution of local Dark Matter shows no indication of homogeneity, but instead closely follows
231
+ the fractal structures of the galaxy distribution.
232
+ These results, taken together, suggest a model of galaxy formation in a D = 2 hierarchical
233
+ IGM according to which all of the matter M0 in a sphere R0 coalesces about a unique center
234
+ so that hierachical symmetry is broken, with (M0, R0) then representing the lower cut-off scales
235
+ of the hierarchy. Given the results of Lelli, McGaugh & Schombert (2016B) to the effect that,
236
+ within any given galaxy, M0 primarily consists of baryonic matter then, given that the resulting
237
+ galactic objects are in equilibrium with the general environment, this model of galaxy formation
238
+ gives a direct derivation of the Baryonic Tully-Fisher Relationship and, consequently, provides
239
+ a natural interpretation of MOND as a first-order descriptor of gravitational dynamics in an
240
+ hierachical cosmos.
241
+ In conclusion, in an hierchical cosmos, the Dark Matter hypothesis and Milgrom’s MOND hy-
242
+ pothesis are seen to be two sides of the same coin.
243
+ References
244
+ Baryshev, Yu V., Sylos Labini, F., Montuori, M., Pietronero, L., astro-ph/9803142
245
+ Hong, S.E., Jeong, D., Hwang, H.S., Kim, J., 2021. Ap. J.; 913, 76
246
+ Lelli, F., McGaugh, SS, Schombert, JM., ApJ., 152, 6, 2016A
247
+ Lelli, F., McGaugh, SS, Schombert, JM., ApJL., 816, L14, 2016B
248
+ 8
249
+
250
+ Milgrom, M., 1983a, Ap. J. 270: 365.
251
+ Milgrom, M., 1983b, Ap. J. 270: 371
252
+ Milgrom, M. 1983c. Ap. J. 270: 384-389
253
+ Mizuno, K., Ishii, J., Kishida, H., Hayamizu, Y., Yasuda, S., Futaba, D., Yumura, M., Hata, K.,
254
+ 2009. PNAS, 106, 15, 6044-6047
255
+ Tekhanovich D.I.I and Baryshev Yu.V., Astro.ph/1610.05206
256
+ A
257
+ Equilibrium at the lower cut-off scales of the hierarchy
258
+ In the cosmos of our experience, galaxies in general appear to be stable and long-lasting struc-
259
+ tures. Since the matter distribution in the D = 2 fractal hierarchy is isotropic (by definition)
260
+ about any arbitrarily chosen centre, then the notional gravitational acceleration imparted to
261
+ a particle at radius R from the centre, and generated by the material contained within R, is
262
+ directed towards the chosen centre and has magnitude given by
263
+ M(R) G
264
+ R2
265
+ = 4πG ΣF ≡ aF, R < ∞.
266
+ (4)
267
+ On this basis, it is clear that the net actual gravitational acceleration imparted to a material
268
+ particle immersed anywhere in the global hierarchy is zero, from which it can be concluded that
269
+ a D = 2 fractal distribution of material is in a state of dynamical equilibrium.
270
+ It follows that:
271
+ • if a finite spherical volume, radius R0, is imagined emptied of all material, then the net
272
+ actual gravitational acceleration of any material particle placed on R0 will be aF directed
273
+ radially outwards from the centre of the empty volume;
274
+ • the empty spherical volume is unstable since all accelerations on R0 are outward. It follows
275
+ that stability requires the volume to be occupied by a stablizing mass, a galaxy say, creating
276
+ a state of zero net radial acceleration on R0. In other words, the equilibrium condition
277
+ g0 ≡ V 2
278
+ 0
279
+ R0
280
+ = aF
281
+ (5)
282
+ must be satisfied.
283
+ 9
284
+
1NE0T4oBgHgl3EQf_gL8/content/tmp_files/load_file.txt ADDED
@@ -0,0 +1,155 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ filepath=/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf,len=154
2
+ page_content='Dark Matter and MOND: Two sides of the same coin?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
3
+ page_content=' D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
4
+ page_content=' F.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
5
+ page_content=' Roscoe (The Open University;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
6
+ page_content=' D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
7
+ page_content='Roscoe@open.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
8
+ page_content='ac.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
9
+ page_content='uk) ORCID: 0000-0003-3561-7425 1 arXiv:2301.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
10
+ page_content='02829v1 [astro-ph.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
11
+ page_content='GA] 7 Jan 2023 Abstract It has recently been reported that the application of convolutional neural-network tech- niques to infer the dark-matter distribution in the local cosmos has revealed how it follows the D ≈ 2 hierarchical distribution of galaxies in the locality, rather than exhibiting the expected homogeneity throughout the IGM.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
12
+ page_content=' Taken at face value, this implies that the Hub- ble Law, observed to be followed on scales which are deep inside the observed hierarchical structures, can no longer be assumed to arise from universal expansion.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
13
+ page_content=' So, if not universal expansion, then what?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
14
+ page_content=' As a possibility, it has been recognized for a considerable time that if the lower cut-off scales of a D ≈ 2 hierarchical cosmos are identified with the scales of a typical galaxy, then gravitational redshift automatically follows the Hubble Law with Hg ≈ 70 km/sec/Mpc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
15
+ page_content=' Inter alia, this suggests a model of galaxy formation in a D ≈ 2 hierarchical IGM in which all of the material M0 within a sphere R0 coalesces about a unique center so that hierachical symmetry is broken on the scale (M0, R0).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
16
+ page_content=' Putting these things together leads unambiguously to the conclusion that, in an hierachical cosmos, the Dark Matter hypothesis and Milgrom’s MOND hypothesis are two sides of the same coin.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
17
+ page_content=' 2 1 Introduction: It is now widely accepted that on scales up to about 200 Mpc galaxies are distributed in a quasi- fractal D ≈ 2 fashion.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
18
+ page_content=' For fairly recent work see Tekhanovich & Baryshev (2016), but many others have contributed over the years.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
19
+ page_content=' It then becomes a point of considerable significance that the Hubble Law is well established on scales that are deep inside the accepted fractal structure of the general galaxy distribution.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
20
+ page_content=' This very interesting circumstance is the primary evidence supporting the idea that the IGM is largely populated by an homogeneous distribution of dark matter on the small scales required for, without homogeneity, the linear nature of Hubble’s Law cannot be understood within the context of universal expansion.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
21
+ page_content=' It is for this reason that the paper of Hong et al (2021) caused so much consternation: specifically, the authors used state-of-the-art convolutional neural-network techniques combined with modern positional and peculiar velocity data to compute and map the local dark matter distribution.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
22
+ page_content=' Against expectation, this distribution is found to trace the hierarchical distribution of galaxies very closely - there is no indication of homogeneity, and hence no indication that the Hubble Law can be understood in terms of universal expansion.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
23
+ page_content=' The only immediately plausible alternative is some form of gravitational redshift: Baryshev et al (1998) point out that inside a D = 2 hierarchical galaxy distribution (with an assumed homogeneous distribution of dark matter) the gravitational part of redshift is also purely linear with distance and cannot be distinguished from the expansion component.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
24
+ page_content=' But if the results of Hong et al (2021) are to be taken at face value, then any contribution to redshift from expansion must manifest itself as a departure from linearity.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
25
+ page_content=' Since such a departure is not observed then, according to the results of Hong et al, there can be no expansion effect at all.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
26
+ page_content=' This line of argument is reinforced by the further observation of Baryshev et al that if the lower cut-off mass and length scales of the hierarchy are identified with the mass and length scales of the typical galaxy, then a gravitational redshift of Hg ≈ 70 km/sec/Mpc is to be ex- pected.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
27
+ page_content=' Inter alia, the foregoing considerations suggest a process of galaxy formation according to which an isolated galactic object can be modelled as a finite bounded spherically symmetric peturbation of the hierarchical IGM (assumed in the first instance to be a mix of baryonic and non-baryonic mass) - this automatically entails that all of the mass M0 within the sphere R0 has coalesced around a unique centre so that fractal symmetry is broken on the scales of (M0, R0).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
28
+ page_content=' 2 Consequences on the lower cut-off scales: From these general considerations we may conclude: 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
29
+ page_content=' The lower cut-off radial and mass scales (M0, R0) must behave according to M0 = 4πR2 0ΣF (1) 3 where ΣF is the mass surface density of the D = 2 hierarchical mass distribution in the local cosmos;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
30
+ page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
31
+ page_content=' Since galaxies in general appear to be stable structures, there must be an equilibrium constraint at the lower cut-off scales of the hierarchy.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
32
+ page_content=' Using simple Newtonian arguments, we show in appendix §A that equilibrium at these lower cut-off scales requires: V 2 0 R0 = aF ≡ 4πGΣF (2) where aF is the characteristic acceleration scale associated with ΣF;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
33
+ page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
34
+ page_content=' The relationship V 4 0 = aF GM0 , (3) which is formally identical to the Baryonic Tully-Fisher Relationship (BTFR), is now de- rived directly by eliminating R0 between (1) and (2).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
35
+ page_content=' It is to be noted that whilst (3) is formally identical to the BTFR of Milgrom’s MOND, it differs fundamentally in the assumption (expressed in the last paragraph of §1) that M0 is an unknown mix of baryonic and non-baryonic mass whereas, by definition, the BTFR asserts that this mass is purely baryonic.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
36
+ page_content=' 3 Empirical support for the BTFR hypothesis 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
37
+ page_content='1 The analysis of Lelli, McGaugh & Schombert (2016B) It has only recently been possible to explore the BTFR hypothesis in a statistically rigorous fashion.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
38
+ page_content=' Specifically, the SPARC sample of Lelli, McGaugh & Schombert (2016A) contains high quality rotation curves and high quality modern surface photometry at 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
39
+ page_content='6 µm for a sample of 175 nearby disk galaxies.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
40
+ page_content=' The high quality of the surface photometry over this sample allowed Lelli, McGaugh & Schombert (2016B) to construct photometric models of baryonic mass distributions in that particular subsample of 118 disks which also had rotation curves extending to flatness, making it ideal for a statistically rigorous testing of the BTFR hypothesis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
41
+ page_content=' Subsequently, the authors used regression analysis techniques to demonstrate how the subsample really does fit the BTFR with very small scatter.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
42
+ page_content=' In this way, they argued that the observed scatter is sufficiently below the instrinsic-scatter expectations of ΛCDM cosmology to present a fundamental difficulty for that cosmology and for the associated idea of dynamically significant quantities of non-baryonic matter in the generality of galaxy disks.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
43
+ page_content=' Since (3) is derived from the hypothesis that galaxies form by coalescing in a stable way out of the D = 2 hierarchical IGM, this result implies that the IGM itself consists primarily of undetected baryonic matter and so, in effect, (3) itself represents a derivation of the hitherto empirical BTFR from a fundamental theoretical position.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
44
+ page_content=' 4 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
45
+ page_content='2 The estimation of (aF, ΣF) The data used by Lelli, McGaugh & Schombert (2016B) is available as an on-line data-sheet giv- ing estimates for the photometrically modelled baryonic masses M0 and flat rotation velocities V0 for the 118 disk galaxies.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
46
+ page_content=' Given this data, an alternative demonstration supporting the BTFR hypothesis is provided by showing how the hypothesis, applied differently to the data, yields a very sharp estimate of the characteristic acceleration parameter aF, thereby demonstrating how, for all practical purposes, its value is identical to that of Milgrom’s critical acceleration parame- ter, a0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
47
+ page_content=' In order to estimate aF ≡ 4πGΣF from this data, we rearrange (3) as V 4 0 GM0 = 4πGΣF ≡ constant and hence form the empirical sample distribution J ≡ � V 4 0i GM0i , i = 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
48
+ page_content='..' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
49
+ page_content='118 � .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
50
+ page_content=' Then, from J, we generate N = 10000 bootstrapped distributions, ˆJi, i = 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
51
+ page_content='.N in the usual way.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
52
+ page_content=' For each ˆJi we then compute its geometric mean, ˆaFi say, to obtain, finally, the distribution AF ≡ (log ˆaFi, i = 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
53
+ page_content='.N) .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
54
+ page_content=' The density distribution of AF is given in figure 1 from which it is clear that the estimate for aF is very tightly constrained around the modal value of 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
55
+ page_content='3 × 10−10 mtrs/sec2 which, for all practical purposes, is identical to Milgrom’s value of MOND’s critical acceleration parameter a0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
56
+ page_content=' This estimate of aF corresponds to ΣF ≈ 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
57
+ page_content='15 kg/mtr2 for the mass surface density of the hierarchical cosmos.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
58
+ page_content=' The grey curve in figure 1 arises from the same analysis but applied to shuffled velocity and mass data.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
59
+ page_content=' It is clear that the signal so powerfully present in the unshuffled data is destroyed by shuffling.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
60
+ page_content=' We conclude that, for all practical purposes the signal for the mass surface density ΣF ≈ 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
61
+ page_content='15 kg/mtr2 in the hierarchical cosmos is real.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
62
+ page_content=' 5 Dotted = 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
63
+ page_content='1 × 10−10 mtrs/sec2 Mode = 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
64
+ page_content='3 × 10−10 mtrs/sec2 Dashed = 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
65
+ page_content='6 × 10−10 mtrs/sec2 0 5 10 15 log(aF) Density Distribution of bootstrapped geometric means Figure 1: Solid black curve = density distribution of log ˆaF.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
66
+ page_content=' Solid grey curve arises when velocity and mass data are shuffled with respect to each other.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
67
+ page_content=' The signal represented by the black curve is destroyed on the shuffled data.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
68
+ page_content=' 4 Full circle to MOND Taking the results of Hong et al (2021) at face value,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
69
+ page_content=' together with the results of Lelli,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
70
+ page_content=' McGaugh & Schombert (2016B) and observations of Baryshev et al (1998),' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
71
+ page_content=' it has been shown that the Dark Matter of modern astrophysics,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
72
+ page_content=' rather than being an homogeneous distribution of non-baryonic matter,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
73
+ page_content=' can reasonably be identified as a D = 2 hierarchical distribution of undetected baryonic matter which,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
74
+ page_content=' as the BTFR (3) shows,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
75
+ page_content=' provides exactly the dynamical support for galaxies that the original non-baryonic Dark Matter hypothesis was formulated for in the first place.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
76
+ page_content=' 6 However, hitherto the prominence of the originally empirically derived BTFR has rested en- tirely upon the fact that it is central to the architecture of MOND Milgrom (1983a,b,c).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
77
+ page_content=' To see this, (3) gives directly V 2 0 R0 = √aF GM0 R0 which, in effect, MOND extrapolates by hypothesis to give g = √aF gN, where gN ≡ GM0/R2, as the effective gravitational force for all R ≥ R0 .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
78
+ page_content=' Consequently, MOND automatically receives the direct interpretation as a first-order descriptor of gravitational dynamics in a D = 2 hierarchical IGM.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
79
+ page_content=' In short, in an hierachical cosmos, the Dark Matter hypothesis and Milgrom’s MOND hypothesis are two sides of the same coin.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
80
+ page_content=' 5 The nature of baryonic Dark Matter?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
81
+ page_content=' We have argued that the Dark Matter of the IGM is distributed in a D = 2 hierarchy and consists of undetected baryonic material.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
82
+ page_content=' So, the immediate question is: how can such a distribution of baryonic material remain undetected?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
83
+ page_content=' There are three potential strands to the answer, easily conceived to be acting in concert.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
84
+ page_content=' 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
85
+ page_content='1 Conventional possibilities Given that the IGM forms a D = 2 hierarchy, then its volume density in a spherical volume of radius R tends to zero as R → ∞, making its detection at large radii intrinsically difficult.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
86
+ page_content=' Furthermore, it can reasonably be assumed that the IGM is at least close to being in thermal equilibrium with the general background, again making its detection against the background also intrinsically difficult.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
87
+ page_content=' 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
88
+ page_content='2 Unconventional possibilities Until recently, it has always been assumed that there is no such thing in nature as a perfect, or near perfect, blackbody absorber - the reason being that no such thing had ever been observed.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
89
+ page_content=' However, Mizuno et al (2009) showed how to fabricate, from agglomerations of single-walled carbon nanotubes (SWCNTs), material distributions having specific bulk statistics which act as near-perfect blackbody absorbers (emissivity > 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
90
+ page_content='98) across a very wide range of incident wavelengths from UV at 200nm to the far IR at 200µm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
91
+ page_content=' This behaviour has been shown to be independent of the specific properties of the individual SWCNTs, but is rather a consequence of the bulk statistical characteristics of the fabricated SWCNT distributions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
92
+ page_content=' We know that many allotropes of carbon exist in interstellar space and these must to some extent be blown into the IGM from the generality of galactic interiors.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
93
+ page_content=' It is a short step to visualizing the existence of clouds of SWCNTs dispersed throughout the hierachical 7 IGM containing sub-populations which, when viewed in projection along any given line of sight, possess the bulk statistical characteristics required to mimic the properties of the fabricated SWCNT distributions of Mizuno et al (2009).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
94
+ page_content=' In this way, it is possible to conceive how SWCNT clouds within the IGM have the potential to act as ‘dispersed near-perfect blackbody objects’ making them virtually undetectable.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
95
+ page_content=' 6 Summary and conclusions There is general agreement that the distribution of galaxies in particular is quasi-fractal D ≈ 2 out to about 200 Mpc and Baryshev et al (1998) has pointed out that gravitational redshift in such an hierarchical cosmos will follow the Hubble Law.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
96
+ page_content=' Furthermore, these authors point out that if the lower cut-off scales of the hierarchy are identified with the mass and radial scales of the typical galaxy, then Hg ≈ 70 km/sec/Mpc is to be expected.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
97
+ page_content=' Notwithstanding the hierarchical distribution of galaxies in the local cosmos, the conventional view holds that the IGM itself consists of non-baryonic Dark Matter, the assumed homogeneous distribution of which makes the Hubble Law a consequence of universal expansion.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
98
+ page_content=' It is for this reason that the paper of Hong et al (2021), which computes and maps the distribution of Dark Matter in the local cosmos, caused so much consternation: specifically, they report that the distribution of local Dark Matter shows no indication of homogeneity, but instead closely follows the fractal structures of the galaxy distribution.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
99
+ page_content=' These results, taken together, suggest a model of galaxy formation in a D = 2 hierarchical IGM according to which all of the matter M0 in a sphere R0 coalesces about a unique center so that hierachical symmetry is broken, with (M0, R0) then representing the lower cut-off scales of the hierarchy.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
100
+ page_content=' Given the results of Lelli, McGaugh & Schombert (2016B) to the effect that, within any given galaxy, M0 primarily consists of baryonic matter then, given that the resulting galactic objects are in equilibrium with the general environment, this model of galaxy formation gives a direct derivation of the Baryonic Tully-Fisher Relationship and, consequently, provides a natural interpretation of MOND as a first-order descriptor of gravitational dynamics in an hierachical cosmos.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
101
+ page_content=' In conclusion, in an hierchical cosmos, the Dark Matter hypothesis and Milgrom’s MOND hy- pothesis are seen to be two sides of the same coin.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
102
+ page_content=' References Baryshev, Yu V.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
103
+ page_content=', Sylos Labini, F.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
104
+ page_content=', Montuori, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
105
+ page_content=', Pietronero, L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
106
+ page_content=', astro-ph/9803142 Hong, S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
107
+ page_content='E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
108
+ page_content=', Jeong, D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
109
+ page_content=', Hwang, H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
110
+ page_content='S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
111
+ page_content=', Kim, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
112
+ page_content=', 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
113
+ page_content=' Ap.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
114
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
115
+ page_content=';' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
116
+ page_content=' 913, 76 Lelli, F.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
117
+ page_content=', McGaugh, SS, Schombert, JM.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
118
+ page_content=', ApJ.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
119
+ page_content=', 152, 6, 2016A Lelli, F.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
120
+ page_content=', McGaugh, SS, Schombert, JM.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
121
+ page_content=', ApJL.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
122
+ page_content=', 816, L14, 2016B 8 Milgrom, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
123
+ page_content=', 1983a, Ap.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
124
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
125
+ page_content=' 270: 365.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
126
+ page_content=' Milgrom, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
127
+ page_content=', 1983b, Ap.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
128
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
129
+ page_content=' 270: 371 Milgrom, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
130
+ page_content=' 1983c.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
131
+ page_content=' Ap.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
132
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
133
+ page_content=' 270: 384-389 Mizuno, K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
134
+ page_content=', Ishii, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
135
+ page_content=', Kishida, H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
136
+ page_content=', Hayamizu, Y.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
137
+ page_content=', Yasuda, S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
138
+ page_content=', Futaba, D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
139
+ page_content=', Yumura, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
140
+ page_content=', Hata, K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
141
+ page_content=', 2009.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
142
+ page_content=' PNAS, 106, 15, 6044-6047 Tekhanovich D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
143
+ page_content='I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
144
+ page_content='I and Baryshev Yu.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
145
+ page_content='V.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
146
+ page_content=', Astro.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
147
+ page_content='ph/1610.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
148
+ page_content='05206 A Equilibrium at the lower cut-off scales of the hierarchy In the cosmos of our experience, galaxies in general appear to be stable and long-lasting struc- tures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
149
+ page_content=' Since the matter distribution in the D = 2 fractal hierarchy is isotropic (by definition) about any arbitrarily chosen centre, then the notional gravitational acceleration imparted to a particle at radius R from the centre, and generated by the material contained within R, is directed towards the chosen centre and has magnitude given by M(R) G R2 = 4πG ΣF ≡ aF, R < ∞.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
150
+ page_content=' (4) On this basis, it is clear that the net actual gravitational acceleration imparted to a material particle immersed anywhere in the global hierarchy is zero, from which it can be concluded that a D = 2 fractal distribution of material is in a state of dynamical equilibrium.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
151
+ page_content=' It follows that: if a finite spherical volume, radius R0, is imagined emptied of all material, then the net actual gravitational acceleration of any material particle placed on R0 will be aF directed radially outwards from the centre of the empty volume;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
152
+ page_content=' the empty spherical volume is unstable since all accelerations on R0 are outward.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
153
+ page_content=' It follows that stability requires the volume to be occupied by a stablizing mass, a galaxy say, creating a state of zero net radial acceleration on R0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
154
+ page_content=' In other words, the equilibrium condition g0 ≡ V 2 0 R0 = aF (5) must be satisfied.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
155
+ page_content=' 9' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/1NE0T4oBgHgl3EQf_gL8/content/2301.02829v1.pdf'}
1tAyT4oBgHgl3EQfPfba/content/tmp_files/2301.00027v1.pdf.txt ADDED
The diff for this file is too large to render. See raw diff
 
1tAyT4oBgHgl3EQfPfba/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
2tAyT4oBgHgl3EQfPvai/content/tmp_files/2301.00031v1.pdf.txt ADDED
@@ -0,0 +1,781 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+
3
+
4
+
5
+
6
+
7
+ ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII
8
+ WEB: https://www.ankarakongresi.org
9
+ E-MAIL: [email protected]
10
+ 1657
11
+ PREDICTING THE STUDENTS INVOLVEMENTS AND IT’S IMPACTS ON
12
+ LEARNING OUTCOMES THROUGH ONLINE EDUCATION DURING COVID-19
13
+
14
+ Muhammad Nadeem
15
+ Computer Science Department, University of the Punjab
16
+
17
+
18
+ Faisal Bukhari
19
+ Data Science Department, University of the Punjab
20
+
21
+ Ali Hussain
22
+ Computer Science Department, University of the Punjab
23
+
24
+ Abstract
25
+ Everybody knows very well about the COVID-19 pandemic, lockdown, and its impacts and
26
+ effects on every field of life, from childhood to senior citizens, from local to global. The
27
+ underlying research study focuses on students' involvement in online classes. This paper
28
+ assesses the effect of the COVID-19 pandemic on the students' participation and involvement
29
+ during online classes compared to the physical classes, cheating behavior, health effects, and
30
+ study styles of the students of diverse degrees and age groups. This research study contributes
31
+ to the real problems and challenges that students faced during online classes during the
32
+ COVID-19 pandemic. The percentages of the students' responses with different color schemes
33
+ shown in Fig. 1, Fig. 2, Fig.3(a), Fig.3(b) and Fig.4 are conveying powerful and meaningful
34
+ insight. These figures and the results given in Table I and Table II indicate that most students
35
+ are not fully involved during online classes due to technical issues, remote distance, etc. We
36
+ applied the Test here because we do not have exact population means. We used ttest_1samp
37
+ with default value 0 to compute the variables' statistics and p-value. These values are minimal
38
+ in favor of rejecting the null or H0 (hypothesis) and accepting the alternate or H1
39
+ (hypothesis). It further means that students' involvement during online classes is severely
40
+ affected.
41
+ Keywords: COVID-19, e-Learning, Students Involvements, Cheating Concerns of Students,
42
+ Class Participation.
43
+
44
+ I. INTRODUCTION
45
+ The primary motivation for selecting this topic is that the quality of education is directly
46
+ proportional to the involvement of the students during the lecture. Firstly, I found it as a
47
+ teacher that many students have left the online lecture physically, but logically they showed
48
+ their status as a present. This problem has multiple issues. The respected teacher cannot be
49
+ confident about the presence of the students physically during online lectures. Secondly, the
50
+ students are facing different issues during online lectures. The impact of these issues is that
51
+ they lose interest in learning during online lectures. This research work is a new study focused
52
+ mainly on the level of student involvement during online lectures. All the countries attacked
53
+ by the villainous COVID-19 virus that has upset each area of life as per economy, from
54
+ producers to consumers [1]. During the Covid19 pandemic, the Education sector was also
55
+ severely impacted. The forceful impact of this virus sent the students and teachers to study
56
+ and teach remotely from face to face system of education. Resultantly, Educational
57
+ institutions are searching for another way to teach and evaluate the students [2]. So to keep
58
+ every student and teacher safe, all the Educational Institutions closed because of the citywide,
59
+ districtwide, and countrywide lockdowns. In such lockup situations, the students and teachers
60
+ cannot interact face-to-face [3].
61
+
62
+
63
+
64
+
65
+
66
+
67
+
68
+
69
+ ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII
70
+ WEB: https://www.ankarakongresi.org
71
+ E-MAIL: [email protected]
72
+ 1658
73
+ To keep the chain of teaching in COVID-19 virus, the World Bank has been actively trying to
74
+ give financial assistance to the underdeveloped or more affected countries. The ultimate goal
75
+ of [4] is to provide basic education rights to every student during this viral disease. As far as
76
+ online learning is concerned, there is much use of technology. This technology-dependent
77
+ way of education becomes a barrier for learners who did not train to use technology [5].
78
+ Similarly, in Pakistan, in 2021, all the educational institutions have closed as the previous
79
+ year due to the severity of COVID-19. Pakistani Ministry of Education and Higher Education
80
+ Commission (HEC) also provides online and distance learning ways to teach the students. [6].
81
+ The HEC provided the design for online policy guidance notes and guidelines for the
82
+ Universities. However, It’s a reality that practical work is not being taught during online
83
+ education. This also demotivated the students, and it made an impact on their involvement in
84
+ online lectures [7]. In addition to the problems mentioned above and issues of students and
85
+ teachers, there are also the problems of admin staff [8].
86
+ Therefore, the teachers are not satisfied with the student’s involvement in online classes
87
+ compared to physical classes.
88
+ In this connection, to find the answers, this study would work on the following research
89
+ objectives:
90
+ • To predict why the students are involved is not as much as physical class.
91
+ • To find why the students are not interested in attending the full online lecture.
92
+ • To discover the issue faced by the students during online lecture.
93
+ • To find the impact of taking lectures in class room with the lecture taking online on the
94
+ students' learning outcomes.
95
+ • To find the family members' realization about their children's online study.
96
+ The outcomes of the research would be necessary for the following concerning levels:
97
+ • Student
98
+ • Teacher
99
+ • Parents
100
+ • Educational Institution
101
+ • Education Ministries
102
+ The most crucial stakeholder in the learning process are teachers, and students are aware of
103
+ the issues and the factors involved as per the student involvement during an online class. The
104
+ parents would also notice the difference in attitude and aptitude to study in the classroom and
105
+ at home via online education. The Educational Institution may send reports to the Ministry of
106
+ Education and HEC based on the outcomes of the student's involvement during an online
107
+ class. In this way, the Ministries can inform the Government to look after the policies to plan
108
+ a different mature online education system or to open the educational institution as soon as
109
+ possible.
110
+
111
+ II. LITERATURE REVIEW
112
+ The impacts of COVID-19 on health, society, and education are highlighted in [9]. The
113
+ researchers divided their research into four different groups: general demographics,
114
+ information about daily online routine, assessment of the learning of online experience and
115
+ level of satisfaction of the students, and evaluation of health due to change in lifestyle.
116
+ Cheating during the exam is one of the main problems. The research work done by [10] on
117
+ cheating shows that an individual's strengths vary according to the achievement settings.
118
+ Their findings also concluded that the cheating rate was higher in educational settings than in
119
+ work areas and in work sites than in sports venues. Study 1 further suggests that the strengths
120
+ of individuals' cheating intentions differ across achievement settings.
121
+
122
+
123
+
124
+
125
+
126
+
127
+
128
+
129
+ ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII
130
+ WEB: https://www.ankarakongresi.org
131
+ E-MAIL: [email protected]
132
+ 1659
133
+ Intentions to cheat were higher in educational settings than in work settings and higher in
134
+ work settings than in sports settings. The outcomes of this research [11] concluded that the
135
+ online examination during COVID-19 increased the cheating ratio, which is unrelated to
136
+ achievement goals. The studies provided different guidelines to the teachers for setting the
137
+ questions and time duration for online exams. The researchers of [12] highlighted the levels of
138
+ students' stress, depressive symptoms, loneliness, effects of missing social life, and specific
139
+ worries for their undergraduate studies. They also showed extreme crises of the students on
140
+ health and research during lockdown due to COVID-19. The authors discussed that they got
141
+ 212 responses out of 266 from students for the crises suffered. They also recommended
142
+ different plans for teachers and academic institution administrators to develop online events
143
+ so that they can prepare newcomers very well. The research efforts of [13] discover the
144
+ critical problems faced by the students in the present e-learning system. They have also found
145
+ the factors influencing online learning during COVID-19.
146
+ The authors also discussed the impacts of students' willingness to study alone in an e-learning
147
+ environment. In addition, they interviewed 30 students from six Universities and conducted
148
+ meetings with 31 e-learning system experts to find the main problems. They also suggested
149
+ applicable plans for policymakers, developers, designers, and researchers, enabling them to be
150
+ better acquainted with the critical aspects of the e-learning system during the COVID-19
151
+ pandemic. The researchers of [14] have found too much dissatisfaction during the online
152
+ study on the COVID-19 situation. The outcomes of this research concluded that the students
153
+ of the dental study were dissatisfied with the online teaching during COVID-19. The results
154
+ of this research crying that online study is disturbing the student's level of involvement in the
155
+ study very severely. The efforts of the analyses highlighted different aspects of students
156
+ during the online study in the COVID-19 pandemic worldwide. They discussed and evaluated
157
+ severe issues such as technical and economic issues, psychological problems, and students'
158
+ fears about the future. It badly affects the study taste of the students and their pace in the
159
+ learning process. They also offered different plans and suggestions for the policymakers and
160
+ higher authorities to overcome the issues faced by the students and the teachers. The research
161
+ study by the authors of [15] observed and evaluated the impact of the perception of e-learning
162
+ crashes. They discovered its impact on psychological upset in the students during the COVID-
163
+ 19 pandemic. They concluded that fear of academic loss had become the main reason for
164
+ mental upset during the issues of online study in corona disease. They also suggested
165
+ remedies for the policymakers and educational institutions to manage the student's stress
166
+ during the online study. The researchers analyzed different types of challenges faced by the
167
+ students in Pakistani Universities [16]. The main obstacles highlighted are economic,
168
+ technical, lack of skills, family support, etc. They also recommended that the Govt. take a
169
+ severe step to overcome the challenges faced by the students. The outcomes of this research
170
+ work [17] show that the students do not want to study online. The students expressed their
171
+ problems during the survey that they were not prepared and trained for such a learning shift.
172
+ They do not have a non-stop electricity facility and well-equipped information technology-
173
+ based infrastructure at their homes.
174
+
175
+ III. PROBLEM STATEMENT
176
+ To find the effect of the COVID-19 pandemic on the involvement of the students during
177
+ online classes as compared to the physical classes, cheating behavior, health effects, and study
178
+ styles from the students of diverse degrees and age groups.
179
+ Hypothesis:
180
+ H0 = Student’s involvement during online classes is the same as in physical classes.
181
+ H1 = Student’s involvement during online classes is not the same as in physical classes.
182
+
183
+
184
+
185
+
186
+
187
+
188
+
189
+ ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII
190
+ WEB: https://www.ankarakongresi.org
191
+ E-MAIL: [email protected]
192
+ 1660
193
+ Methodology and Data Collection
194
+ The survey methodology used to accomplish this research. Survey is a method for the
195
+ collection of the information for the sample of individuals [18]. The findings of the survey
196
+ analyzed through statistical analysis.
197
+
198
+ • OBJECTIVES OF THE SURVEY
199
+ To analyze the levels of the student’s involvement and its impacts on learning outcomes
200
+ during online lectures during COVID-19.
201
+
202
+ •TARGET POPULATION
203
+ Graduate, Undergraduate and Intermediate students of the Universities and Colleges
204
+
205
+ •DATA TO BE COLLECTED
206
+ A questionnaire developed based on the literature review. Then this questionnaire circulated
207
+ online as much as possible to find the maximum responses from the target population due to
208
+ the COVID-19 situation.
209
+
210
+ •MEASUREMENT `INSTRUMENT'
211
+ The measurement instrument of the required survey is a questionnaire. The questions of this
212
+ questionnaire were
213
+ closed-ended with a Likert scale. The definition of the Likert scale is given below:
214
+ 1. SA (Strongly Agreed)
215
+ 2. A (Agreed)
216
+ 3. U (Undecided),
217
+ 4. D (Disagreed)
218
+ 5. SD (Strongly Disagreed)
219
+ This questionnaire would be distributed through Google docs to make it available to the
220
+ targeted population and to get a maximum number of responses.
221
+
222
+ IV. DESIGN OF RESEARCH STUDY
223
+ An online survey performed using Google online forms. However, the questionnaire of this
224
+ survey consists of the following subsections:
225
+ A. Respondents will be requested to answer their following usual demographics:
226
+ • Age
227
+ • Gender
228
+ • Area of residence
229
+ B. Getting information routine wise online learning during the shift from face to face study to
230
+ online study in colleges/Universities in Pakistan. These information consists of the following:
231
+ • Average time given for online study in hours per day
232
+ • Quality and the problems of the communication medium
233
+ • Actual involvement in virtual lecture same as face to face lecture in physical class
234
+ • Level of interruption by the family members during online study period
235
+ • Attention and focus level from joining to the end of online class.
236
+ • Effects of online learning on Cheating behavior and students involvement to
237
+ C. Evaluation of the experience of the student’s level of involvement in virtual class to find
238
+ the overall students involvement in online lecture.
239
+ D. Evaluation of health during change in learning style from physical class environment
240
+ provided by the College/University to the virtual class environment provided by your parents
241
+ at home and the effects of virtual class on your involvement of class.
242
+
243
+
244
+
245
+
246
+
247
+
248
+
249
+
250
+ ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII
251
+ WEB: https://www.ankarakongresi.org
252
+ E-MAIL: [email protected]
253
+ 1661
254
+ The pictorial survey responses are given below:
255
+
256
+
257
+ Fig. 1. Getting General Info
258
+
259
+
260
+ Fig. 2. Getting General Info
261
+
262
+
263
+ Fig. 3(a). Getting Specific Info
264
+
265
+
266
+
267
+ Section A: Demographics Info
268
+ Chart
269
+ 100%
270
+ 80%
271
+ 60%
272
+ 40%
273
+ 20%
274
+ 0%
275
+ 6
276
+ 1
277
+ 4
278
+ 4
279
+ 5
280
+ 8
281
+ 5
282
+ 9
283
+ 3
284
+ 2
285
+ 3
286
+ 3
287
+ 3
288
+ 4
289
+ 4
290
+ 4
291
+ 5
292
+ 5
293
+ 5
294
+ Gender
295
+ Age
296
+ Degree_ Level
297
+ Area of ResidenceSection B: Getting General Info Chart
298
+ 100%
299
+ 80%
300
+ 60%
301
+ 40%
302
+ 20%
303
+ 0%
304
+ 3
305
+ 5
306
+ 9
307
+ 2
308
+ 8
309
+ 5
310
+ 8
311
+ 385
312
+ 417
313
+ 1.49
314
+ 3
315
+ 4
316
+ 8
317
+ 3
318
+ 6
319
+ 1
320
+ 1
321
+ T
322
+ 2
323
+ 3
324
+ 3
325
+ 5
326
+ 5
327
+ 5
328
+ 6
329
+ Time_Spent_SociaMediaLaptopComputerAvail
330
+ ISmartPhonesAvail
331
+ Class Participation Level
332
+ CheatingConcern
333
+ StudyLevelifdontExam
334
+ I Lack of IT Skills
335
+ BetterOnlineLearnSectionC:GettingSpecificInfo(11-17
336
+ Questions)
337
+ 100%
338
+ 80%
339
+ 60%
340
+ 40%
341
+ 20%
342
+ 0%
343
+ 8
344
+ 6
345
+ OnlineAndOffineEqual
346
+ Technicallssuelmpact
347
+ Economiclssuelmpact
348
+ TeacherVoicelssue
349
+ Lessrsinteraction
350
+ AcademicLossfearinClassParticipation
351
+ LackDeficiencyforNonITSt
352
+
353
+
354
+
355
+
356
+
357
+ ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII
358
+ WEB: https://www.ankarakongresi.org
359
+ E-MAIL: [email protected]
360
+ 1662
361
+ We have created questionnaire. Its soft copy is available at the following link:
362
+ https://docs.google.com/forms/d/1zqnXC9EXRXjmNL7VX2FP4hh6OL0NVu-
363
+ C_w8QZOFxRsc/edit )
364
+ We have collected 623 responses from different level of degree students.
365
+
366
+
367
+ Fig. 3(b). Getting Specific Info
368
+
369
+
370
+ Fig. 4. Getting Health Issues info
371
+
372
+ V. EXPEIRMENTAL RESULTS
373
+ The means and standard deviation of all the variables as per questionnaire are given below:
374
+
375
+
376
+
377
+ SectionC:GettingSpecificInfo(1-10Questions)
378
+ 100%
379
+ 80%
380
+ 60%
381
+ 40%
382
+ 20%
383
+ 0%
384
+ 8
385
+ 5
386
+ 2
387
+ 5
388
+ BetterTimeUtilization
389
+ CheatingBehavior
390
+ Umwilingness_of_ResponsibilityStudentsHesitancylmpact
391
+ TechDificultylmpact
392
+ HaveNetAccess
393
+ HaveElectricSuppy
394
+ InteractionWihTeacher
395
+ ClassParticipationChance
396
+ AttensionAndFocusDisturbSection D: Getting Health Issues Info
397
+ Chart
398
+ 100%
399
+ 80%
400
+ 60%
401
+ 40%
402
+ 20%
403
+ 0%
404
+ 4
405
+ 8
406
+ 2
407
+ 3
408
+ 80
409
+ 365
410
+ 393
411
+ 2
412
+ 4
413
+ 3
414
+ 89
415
+ 9
416
+ 3
417
+ 3
418
+ 4
419
+ 5
420
+ 5
421
+ 5
422
+ 5
423
+ 6
424
+ FitnessOfAttensionLonelinessEffectAnxietyLevelPsycholmpacts
425
+
426
+
427
+
428
+
429
+
430
+ ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII
431
+ WEB: https://www.ankarakongresi.org
432
+ E-MAIL: [email protected]
433
+ 1663
434
+ TABLE I.
435
+ S.#
436
+ Variable
437
+ Value
438
+ Mean of all the variables
439
+ SECTION A: DEMOGRAPHICS INFO
440
+ 1
441
+ Gender
442
+ 1.400000
443
+ 2
444
+ Age
445
+ 2.028571
446
+ 3
447
+ Degree_Level
448
+ 2.257143
449
+ 4
450
+ Area_of_Residence
451
+ 1.628571
452
+ SECTION B: GETTING GENERAL INFO
453
+ 1
454
+ Time_Spent_ SociaMedia
455
+ 2.457143
456
+ 2
457
+ LaptopComputerAvail
458
+ 1.771429
459
+ 3
460
+ SmartPhonesAvail
461
+ 1.571429
462
+ 4
463
+ Class_Participation_Level
464
+ 2.885714
465
+ 5
466
+ CheatingConcern
467
+ 1.942857
468
+ 6
469
+ StudyLevelIfdontExam
470
+ 3.028571
471
+ 7
472
+ Lack_of_IT_Skills
473
+ 2.485714
474
+ 8
475
+ BetterOnlineLearn
476
+ 3.600000
477
+ SECTION C: GETTING SPECIFIC INFO
478
+ 1
479
+ BetterTimeUtilization
480
+ 3.342857
481
+ 2
482
+ CheatingBehavior
483
+ 2.285714
484
+ 3
485
+ Unwilingness_of_Responsibility
486
+ 2.114286
487
+ 4
488
+ StudentsHesitancyImpact
489
+ 2.371429
490
+ 5
491
+ TechDifficultyImpact
492
+ 2.200000
493
+ 6
494
+ HaveNetAccess
495
+ 2.514286
496
+ 7
497
+ HaveElectricSupply
498
+ 3.000000
499
+ 8
500
+ InteractionWihTeacher
501
+ 3.085714
502
+ 9
503
+ ClassParticipationChance
504
+ 3.057143
505
+ 10
506
+ AttensionAndFocusDisturb
507
+ 2.342857
508
+ 11
509
+ OnlineAndOfflineEqual
510
+ 3.800000
511
+ 12
512
+ TechnicalIssueImpact
513
+ 1.857143
514
+ 13
515
+ EconomicIssueImpact
516
+ 2.000000
517
+ 14
518
+ TeacherVoiceIssue
519
+ 1.971429
520
+ 15
521
+ LessTSInteraction
522
+ 1.857143
523
+ 16
524
+ AcademicLossFearinClassParticipation
525
+ 2.028571
526
+ 16
527
+ AcademicLossFearinClassParticipation
528
+ 0.970588
529
+ 17
530
+ LackDeficiencyforNonITSt
531
+ 0.747240
532
+ Standard Deviation of all the variables
533
+ SECTION A: DEMOGRAPHICS INFO
534
+ 1
535
+ Gender
536
+ 0.489898
537
+ 2
538
+ Age
539
+ 0.376883
540
+ 3
541
+ Degree_Level
542
+ 0.552545
543
+ 4
544
+ Area_of_Residence
545
+ 0.483187
546
+
547
+
548
+
549
+
550
+
551
+
552
+
553
+
554
+
555
+ ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII
556
+ WEB: https://www.ankarakongresi.org
557
+ E-MAIL: [email protected]
558
+ 1664
559
+ SECTION B: GETTING GENERAL INFO
560
+ 1
561
+ Time_Spent_ SociaMedia
562
+ 1.078169
563
+ 2
564
+ LaptopComputerAvail
565
+ 0.897161
566
+ 3
567
+ SmartPhonesAvail
568
+ 0.766652
569
+ 4
570
+ Class_Participation_Level
571
+ 1.259738
572
+ 5
573
+ CheatingConcern
574
+ 1.093954
575
+ 6
576
+ StudyLevelIfdontExam
577
+ 1.502107
578
+ 7
579
+ Lack_of_IT_Skills
580
+ 1.105090
581
+ 8
582
+ BetterOnlineLearn
583
+ 1.515633
584
+ SECTION C: GETTING SPECIFIC INFO
585
+ 1
586
+ BetterTimeUtilization
587
+ 1.413059
588
+ 2
589
+ CheatingBehavior
590
+ 1.110249
591
+ 3
592
+ Unwilingness_of_Responsibility
593
+ 1.259738
594
+ 4
595
+ StudentsHesitancyImpact
596
+ 1.332789
597
+ 5
598
+ TechDifficultyImpact
599
+ 0.979796
600
+ 6
601
+ HaveNetAccess
602
+ 1.273273
603
+ 7
604
+ HaveElectricSupply
605
+ 1.309307
606
+ 8
607
+ InteractionWihTeacher
608
+ 1.295518
609
+ 9
610
+ ClassParticipationChance
611
+ 1.286032
612
+ 10
613
+ AttensionAndFocusDisturb
614
+ 1.392692
615
+ 11
616
+ OnlineAndOfflineEqual
617
+ 1.214202
618
+ 12
619
+ TechnicalIssueImpact
620
+ 0.797957
621
+ 13
622
+ EconomicIssueImpact
623
+ 0.956183
624
+ 3
625
+ TeacherVoiceIssue
626
+ 1.027777
627
+ 4
628
+ LessTSInteraction
629
+ 1.045886
630
+
631
+ TABLE II:
632
+ TTest Outcomes
633
+ Ttest_1sampResult(statistic=array([16.663333 , 31.38507589,
634
+ 23.81939622, 19.65311057, 13.28871279,
635
+ 11.51311097, 11.95187108, 13.35711613, 10.35574591, 11.7564528 ,
636
+ 13.11574349, 13.84994208, 13.79421828, 12.0044142 , 9.78640192,
637
+ 10.375 , 13.09261879, 11.51416659, 13.36038922, 13.88838218,
638
+ 13.86128572, 9.80912102, 18.24871239, 13.57080199, 12.19631092,
639
+ 11.18462458, 10.35381536, 12.18694645, 13.15416906, 11.9272551 ,
640
+ 11.34226868, 10.64348064, 11.2720409 ]), pvalue=array([6.29551067e-
641
+ 18, 1.08636299e-26, 8.60469002e-23, 3.85366635e-20,
642
+ 5.07753770e-15, 2.80770597e-13, 1.00580609e-13, 4.38220401e-15,
643
+ 4.72979440e-12, 1.58421613e-13, 7.38588068e-15, 1.53985258e-15,
644
+ 1.73086219e-15, 8.90861219e-14, 2.02190243e-11, 4.50637008e-12,
645
+ 7.76732461e-15, 2.80069994e-13, 4.35148773e-15, 1.42079500e-15,
646
+ 1.50369222e-15, 1.90647656e-11, 3.87615125e-19, 2.77545554e-15,
647
+ 5.73530395e-14, 6.15085107e-13, 4.75281129e-12, 5.85928814e-14,
648
+ 6.79392859e-15, 1.06477004e-13, 4.21455437e-13, 2.30654437e-12,
649
+ 4.98568395e-13]))
650
+
651
+
652
+
653
+
654
+
655
+
656
+
657
+
658
+
659
+
660
+ ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII
661
+ WEB: https://www.ankarakongresi.org
662
+ E-MAIL: [email protected]
663
+ 1665
664
+ VIII. CONCLUSTION
665
+ To evaluate and find the correctness and applicability of the hypothesis as per the problem
666
+ statement, we used an online survey approach using Google docs. According to the
667
+ percentages of survey responses given in Fig.1 Fig.2, Fig.3(a), Fig.3(b) and Fig. 4,
668
+ availability of Laptop/Computer at student homes was 85.3% and smart phones was 87.5%.
669
+ Time spent on social media during the online lecture was 71%. The level of Class
670
+ participation was 49.6%. The concern of students cheating during the online exam was
671
+ 70.8%—level of cheating behavior to ignore interest in online due to online exams
672
+ encouraged by 60.8% of students. The student's unwillingness was found at 73.2%. Impacts
673
+ of technical issues during online classes were 84.4% . The pace of the teacher's voice due to
674
+ the Net problem was discovered at 79.5% and Impacts of less interaction of teacher-student
675
+ found to be 78%. As per Fig. 4, psychological impacts on learning participation during online
676
+ classes were discovered at 73.5% , the stress of loneliness affects students' level of
677
+ involvement was 68.5% and the anxiety levels disturb students' level of motivation by 77.9%.
678
+ As per the above experiments, the means and standard deviations are given in Table I and
679
+ Table II above. Most of the high values of means shows that much percentage of the students
680
+ are not fully involved during online lecture. Similarly, most of the values of standard
681
+ deviations are far from zero. It shows that data points are far from the mean. We applied the
682
+ test here because we don't have actual population means. We used ttest_1samp (Dataset
683
+ [:35],0) with a default value of 0 to compute the variables' statistics and p-value. The results
684
+ of this test are provided in Table II. These values are minimal, which is in favor of rejecting
685
+ the null or H0 hypothesis and accepting the alternate or H1 hypothesis. It further means that
686
+ students' involvement during online classes is severely affected.
687
+
688
+ ACKNOWLEDGMENT
689
+ The authors are very grateful to the management of server room of Faculty of Computing and
690
+ Information Technology (FCIT), University of the Punjab to forward our questionnaire to the
691
+ students for the responses. We are also thankful to the Students of Undergraduates and
692
+ Gradates students of FCIT for the warm participation and sincere responses during the survey
693
+ of this research study.
694
+
695
+ REFERENCES
696
+ [1].
697
+ Fernando, R., “The COVID-19 Pandemic: A call for a reality check.”,
698
+ published in Galle Medical Journal, 25 (1). 2020.
699
+ [2].
700
+ Myers, A., “After COVID-19: Recalibrating the American educational
701
+ system”, Retrieved from https://hub.jhu.edu/2020/04/07/bob-balfanz-education-reform-covid-
702
+ 19/, accessed 2021.
703
+ [3].
704
+ Tam, G., & El-Azar, D. (2020), “3 ways the coronavirus pandemic could
705
+ reshape education”, Retrieved from https://www.weforum.org/agenda/2021/05/consumer-
706
+ demand-covid-19-recovery/ published in 2020 but accessed 2021.
707
+ [4].
708
+ World Bank, “How countries are using edtech (including online learning,
709
+ radio, television, texting) to support access to remote learning during the COVID-19
710
+ pandemic.”,
711
+ Retrieved
712
+ from
713
+ https://www.worldbank.org/en/topic/edutech/brief/how-
714
+ countries-are-using-edtech-to-support-remote-learning-during-the-covid-19-pandemic
715
+ ,
716
+ published in 2020 but accessed 2021.
717
+ [5].
718
+ Ractham, P., & Chen, C., “Promoting the use of online social technology as a
719
+ case-based learning tool”, published in Journal of Information Systems Education, 24 (4),
720
+ 2019.
721
+
722
+
723
+
724
+
725
+
726
+
727
+
728
+
729
+ ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII
730
+ WEB: https://www.ankarakongresi.org
731
+ E-MAIL: [email protected]
732
+ 1666
733
+ [6].
734
+ Noor Ul Ain, “Is online education the new future of Pakistan?” Retrieved from
735
+ https://dailytimes.com.pk/579663/blessing-in-disguise-is-online-education-the-new-future-of-
736
+ pakistan/ , published in 2020 but accessed 2021.
737
+ [7].
738
+ Noor Ul Ain Ali., “Students disappointed with online teaching system amid
739
+ COVID-19”. Retrieved from https://dailytimes.com.pk/587446/students-disappointed-with-
740
+ online-teachingsystem-amid-covid-19/, published in 2020 but accessed 2021.
741
+ [8].
742
+ Kebritchi, M., Lipschuetz, A., & Santiague, L., “Issues and challenges for
743
+ teaching successful online courses in higher education: A literature review”, Journal of
744
+ Educational Technology Systems, 46 (1), 4–29, 2017.
745
+ [9].
746
+ Kunal Chaturvedi, Dinesh Kumar Vishwakarma, Nidhi Singh, “ COVID-19
747
+ and its impact on education, social life and mental health of students: A survey”, published in
748
+ Children and Youth Services Review 121 (2021) 105866, pp. 1-6.
749
+ [10]. Nico W. Van Yperen, Melvyn R.W. Hamstra and Marloes van der Klauw, “ To
750
+ Win, or Not to Lose, At Any Cost: The Impact of Achievement Goals on Cheating”,
751
+ published in British Journal of Management, Vol. 22, S5–S15 (2011)
752
+ [11]. Lia M. Daniels, Lauren D, Goeganx Patti C. Parker, “ The impact of COVID
753
+ 19 triggered changes to instruction and assessment on university students’ self reported
754
+ motivation, engagement and perceptions”, published in Social Psychology of Education
755
+ (2021) 24:299–318, Springer.
756
+ [12]. Timon ElmerID, Kieran Mepham, Christoph Stadtfeld, “Students under
757
+ lockdown: Comparisons of students’ social networks and mental health before and during the
758
+ COVID-19 crisis in Switzerland”, published in PLOS ONE, pp: 1-22, 2020.
759
+ [13]. Mohammed Amin Almaiah, Ahmad Al-Khasawneh and Ahmad Althunibat, “
760
+ Exploring the critical challenges and factors influencing the E-learning system usage during
761
+ COVID-19 pandemic”, published in Journal of Education and Information Technologies,
762
+ Springer.
763
+ [14]. Huma Sarwar, Hira Akhtar, Meshal Muhammad Naeem, Javeria Ali Khan,
764
+ Khadija Waraich, Sumaiya Shabbir, Arshad Hasan and Zohaib Khurshid, “ Self-Reported
765
+ Effectiveness of e-Learning Classes during COVID-19 Pandemic: A Nation-Wide Survey of
766
+ Pakistani Undergraduate Dentistry Students”, published in European Journal of Dentistry,
767
+ 2020;14(suppl S1):S34–S43.
768
+ [15]. Aqsa Arshad, Madiha Afzal, * Dr. Muhammad Sabboor Hussain, “ Sudden
769
+ Switch to Post-COVID-19 Online Classes and Cognitive Transformation of ESL Learners:
770
+ Critical Analysis of Discourse of Fear”, published in Research Journal of Social Sciences &
771
+ Economics Review, Vol. 1, Issue 3, 2020, PP: 188-199.
772
+ [16]. Najmul Hasan, Yukun Bao, “ Impact of “e-Learning crack-up” perception on
773
+ psychological distress among college students during COVID-19 pandemic: A mediating role
774
+ of “fear of academic year loss”, published in Children and Youth Services Review 118 (2020)
775
+ PP: 1-10.
776
+ [17]. Muhammad Anwar, Anwar Khan, Khalid Sultan, “The barriers and challenges
777
+ faced by students in online education during covid-19 pandemic in Pakistan”, published in
778
+ Gomal University Journal of Research, Volume 36, Issue 1, JUNE, 2020. PP. 52-62.
779
+ [18]. Brochure, what is a survey? Bill Kalsbeek, 1995 publications officer, ASA
780
+ section on survey research methods, 1995.
781
+
2tAyT4oBgHgl3EQfPvai/content/tmp_files/load_file.txt ADDED
@@ -0,0 +1,407 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ filepath=/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf,len=406
2
+ page_content='ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII WEB: https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
3
+ page_content='ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
4
+ page_content='org E-MAIL: bilgi@ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
5
+ page_content='org 1657 PREDICTING THE STUDENTS INVOLVEMENTS AND IT’S IMPACTS ON LEARNING OUTCOMES THROUGH ONLINE EDUCATION DURING COVID-19 Muhammad Nadeem Computer Science Department,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
6
+ page_content=' University of the Punjab Faisal Bukhari Data Science Department,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
7
+ page_content=' University of the Punjab Ali Hussain Computer Science Department,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
8
+ page_content=' University of the Punjab Abstract Everybody knows very well about the COVID-19 pandemic,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
9
+ page_content=' lockdown,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
10
+ page_content=' and its impacts and effects on every field of life,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
11
+ page_content=' from childhood to senior citizens,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
12
+ page_content=' from local to global.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
13
+ page_content=" The underlying research study focuses on students' involvement in online classes." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
14
+ page_content=" This paper assesses the effect of the COVID-19 pandemic on the students' participation and involvement during online classes compared to the physical classes, cheating behavior, health effects, and study styles of the students of diverse degrees and age groups." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
15
+ page_content=' This research study contributes to the real problems and challenges that students faced during online classes during the COVID-19 pandemic.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
16
+ page_content=" The percentages of the students' responses with different color schemes shown in Fig." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
17
+ page_content=' 1, Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
18
+ page_content=' 2, Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
19
+ page_content='3(a), Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
20
+ page_content='3(b) and Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
21
+ page_content='4 are conveying powerful and meaningful insight.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
22
+ page_content=' These figures and the results given in Table I and Table II indicate that most students are not fully involved during online classes due to technical issues, remote distance, etc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
23
+ page_content=' We applied the Test here because we do not have exact population means.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
24
+ page_content=" We used ttest_1samp with default value 0 to compute the variables' statistics and p-value." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
25
+ page_content=' These values are minimal in favor of rejecting the null or H0 (hypothesis) and accepting the alternate or H1 (hypothesis).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
26
+ page_content=" It further means that students' involvement during online classes is severely affected." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
27
+ page_content=' Keywords: COVID-19, e-Learning, Students Involvements, Cheating Concerns of Students, Class Participation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
28
+ page_content=' I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
29
+ page_content=' INTRODUCTION The primary motivation for selecting this topic is that the quality of education is directly proportional to the involvement of the students during the lecture.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
30
+ page_content=' Firstly, I found it as a teacher that many students have left the online lecture physically, but logically they showed their status as a present.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
31
+ page_content=' This problem has multiple issues.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
32
+ page_content=' The respected teacher cannot be confident about the presence of the students physically during online lectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
33
+ page_content=' Secondly, the students are facing different issues during online lectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
34
+ page_content=' The impact of these issues is that they lose interest in learning during online lectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
35
+ page_content=' This research work is a new study focused mainly on the level of student involvement during online lectures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
36
+ page_content=' All the countries attacked by the villainous COVID-19 virus that has upset each area of life as per economy, from producers to consumers [1].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
37
+ page_content=' During the Covid19 pandemic, the Education sector was also severely impacted.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
38
+ page_content=' The forceful impact of this virus sent the students and teachers to study and teach remotely from face to face system of education.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
39
+ page_content=' Resultantly, Educational institutions are searching for another way to teach and evaluate the students [2].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
40
+ page_content=' So to keep every student and teacher safe, all the Educational Institutions closed because of the citywide, districtwide, and countrywide lockdowns.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
41
+ page_content=' In such lockup situations, the students and teachers cannot interact face-to-face [3].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
42
+ page_content=' ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII WEB: https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
43
+ page_content='ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
44
+ page_content='org E-MAIL: bilgi@ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
45
+ page_content='org 1658 To keep the chain of teaching in COVID-19 virus, the World Bank has been actively trying to give financial assistance to the underdeveloped or more affected countries.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
46
+ page_content=' The ultimate goal of [4] is to provide basic education rights to every student during this viral disease.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
47
+ page_content=' As far as online learning is concerned, there is much use of technology.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
48
+ page_content=' This technology-dependent way of education becomes a barrier for learners who did not train to use technology [5].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
49
+ page_content=' Similarly, in Pakistan, in 2021, all the educational institutions have closed as the previous year due to the severity of COVID-19.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
50
+ page_content=' Pakistani Ministry of Education and Higher Education Commission (HEC) also provides online and distance learning ways to teach the students.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
51
+ page_content=' [6].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
52
+ page_content=' The HEC provided the design for online policy guidance notes and guidelines for the Universities.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
53
+ page_content=' However, It’s a reality that practical work is not being taught during online education.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
54
+ page_content=' This also demotivated the students, and it made an impact on their involvement in online lectures [7].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
55
+ page_content=' In addition to the problems mentioned above and issues of students and teachers, there are also the problems of admin staff [8].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
56
+ page_content=' Therefore, the teachers are not satisfied with the student’s involvement in online classes compared to physical classes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
57
+ page_content=' In this connection, to find the answers, this study would work on the following research objectives: • To predict why the students are involved is not as much as physical class.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
58
+ page_content=' • To find why the students are not interested in attending the full online lecture.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
59
+ page_content=' • To discover the issue faced by the students during online lecture.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
60
+ page_content=" • To find the impact of taking lectures in class room with the lecture taking online on the students' learning outcomes." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
61
+ page_content=" • To find the family members' realization about their children's online study." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
62
+ page_content=' The outcomes of the research would be necessary for the following concerning levels: • Student • Teacher • Parents • Educational Institution • Education Ministries The most crucial stakeholder in the learning process are teachers, and students are aware of the issues and the factors involved as per the student involvement during an online class.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
63
+ page_content=' The parents would also notice the difference in attitude and aptitude to study in the classroom and at home via online education.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
64
+ page_content=" The Educational Institution may send reports to the Ministry of Education and HEC based on the outcomes of the student's involvement during an online class." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
65
+ page_content=' In this way, the Ministries can inform the Government to look after the policies to plan a different mature online education system or to open the educational institution as soon as possible.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
66
+ page_content=' II.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
67
+ page_content=' LITERATURE REVIEW The impacts of COVID-19 on health, society, and education are highlighted in [9].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
68
+ page_content=' The researchers divided their research into four different groups: general demographics, information about daily online routine, assessment of the learning of online experience and level of satisfaction of the students, and evaluation of health due to change in lifestyle.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
69
+ page_content=' Cheating during the exam is one of the main problems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
70
+ page_content=" The research work done by [10] on cheating shows that an individual's strengths vary according to the achievement settings." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
71
+ page_content=' Their findings also concluded that the cheating rate was higher in educational settings than in work areas and in work sites than in sports venues.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
72
+ page_content=" Study 1 further suggests that the strengths of individuals' cheating intentions differ across achievement settings." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
73
+ page_content=' ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII WEB: https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
74
+ page_content='ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
75
+ page_content='org E-MAIL: bilgi@ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
76
+ page_content='org 1659 Intentions to cheat were higher in educational settings than in work settings and higher in work settings than in sports settings.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
77
+ page_content=' The outcomes of this research [11] concluded that the online examination during COVID-19 increased the cheating ratio, which is unrelated to achievement goals.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
78
+ page_content=' The studies provided different guidelines to the teachers for setting the questions and time duration for online exams.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
79
+ page_content=" The researchers of [12] highlighted the levels of students' stress, depressive symptoms, loneliness, effects of missing social life, and specific worries for their undergraduate studies." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
80
+ page_content=' They also showed extreme crises of the students on health and research during lockdown due to COVID-19.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
81
+ page_content=' The authors discussed that they got 212 responses out of 266 from students for the crises suffered.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
82
+ page_content=' They also recommended different plans for teachers and academic institution administrators to develop online events so that they can prepare newcomers very well.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
83
+ page_content=' The research efforts of [13] discover the critical problems faced by the students in the present e-learning system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
84
+ page_content=' They have also found the factors influencing online learning during COVID-19.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
85
+ page_content=" The authors also discussed the impacts of students' willingness to study alone in an e-learning environment." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
86
+ page_content=' In addition, they interviewed 30 students from six Universities and conducted meetings with 31 e-learning system experts to find the main problems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
87
+ page_content=' They also suggested applicable plans for policymakers, developers, designers, and researchers, enabling them to be better acquainted with the critical aspects of the e-learning system during the COVID-19 pandemic.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
88
+ page_content=' The researchers of [14] have found too much dissatisfaction during the online study on the COVID-19 situation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
89
+ page_content=' The outcomes of this research concluded that the students of the dental study were dissatisfied with the online teaching during COVID-19.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
90
+ page_content=" The results of this research crying that online study is disturbing the student's level of involvement in the study very severely." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
91
+ page_content=' The efforts of the analyses highlighted different aspects of students during the online study in the COVID-19 pandemic worldwide.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
92
+ page_content=" They discussed and evaluated severe issues such as technical and economic issues, psychological problems, and students' fears about the future." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
93
+ page_content=' It badly affects the study taste of the students and their pace in the learning process.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
94
+ page_content=' They also offered different plans and suggestions for the policymakers and higher authorities to overcome the issues faced by the students and the teachers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
95
+ page_content=' The research study by the authors of [15] observed and evaluated the impact of the perception of e-learning crashes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
96
+ page_content=' They discovered its impact on psychological upset in the students during the COVID- 19 pandemic.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
97
+ page_content=' They concluded that fear of academic loss had become the main reason for mental upset during the issues of online study in corona disease.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
98
+ page_content=" They also suggested remedies for the policymakers and educational institutions to manage the student's stress during the online study." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
99
+ page_content=' The researchers analyzed different types of challenges faced by the students in Pakistani Universities [16].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
100
+ page_content=' The main obstacles highlighted are economic, technical, lack of skills, family support, etc.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
101
+ page_content=' They also recommended that the Govt.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
102
+ page_content=' take a severe step to overcome the challenges faced by the students.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
103
+ page_content=' The outcomes of this research work [17] show that the students do not want to study online.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
104
+ page_content=' The students expressed their problems during the survey that they were not prepared and trained for such a learning shift.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
105
+ page_content=' They do not have a non-stop electricity facility and well-equipped information technology- based infrastructure at their homes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
106
+ page_content=' III.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
107
+ page_content=' PROBLEM STATEMENT To find the effect of the COVID-19 pandemic on the involvement of the students during online classes as compared to the physical classes, cheating behavior, health effects, and study styles from the students of diverse degrees and age groups.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
108
+ page_content=' Hypothesis: H0 = Student’s involvement during online classes is the same as in physical classes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
109
+ page_content=' H1 = Student’s involvement during online classes is not the same as in physical classes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
110
+ page_content=' ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII WEB: https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
111
+ page_content='ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
112
+ page_content='org E-MAIL: bilgi@ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
113
+ page_content='org 1660 Methodology and Data Collection The survey methodology used to accomplish this research.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
114
+ page_content=' Survey is a method for the collection of the information for the sample of individuals [18].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
115
+ page_content=' The findings of the survey analyzed through statistical analysis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
116
+ page_content=' OBJECTIVES OF THE SURVEY To analyze the levels of the student’s involvement and its impacts on learning outcomes during online lectures during COVID 19.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
117
+ page_content=' TARGET POPULATION Graduate, Undergraduate and Intermediate students of the Universities and Colleges DATA TO BE COLLECTED A questionnaire developed based on the literature review.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
118
+ page_content=' Then this questionnaire circulated online as much as possible to find the maximum responses from the target population due to the COVID 19 situation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
119
+ page_content=" MEASUREMENT `INSTRUMENT' The measurement instrument of the required survey is a questionnaire." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
120
+ page_content=' The questions of this questionnaire were closed ended with a Likert scale.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
121
+ page_content=' The definition of the Likert scale is given below: 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
122
+ page_content=' SA (Strongly Agreed) 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
123
+ page_content=' A (Agreed) 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
124
+ page_content=' U (Undecided), 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
125
+ page_content=' D (Disagreed) 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
126
+ page_content=' SD (Strongly Disagreed) This questionnaire would be distributed through Google docs to make it available to the targeted population and to get a maximum number of responses.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
127
+ page_content=' IV.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
128
+ page_content=' DESIGN OF RESEARCH STUDY An online survey performed using Google online forms.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
129
+ page_content=' However, the questionnaire of this survey consists of the following subsections: A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
130
+ page_content=' Respondents will be requested to answer their following usual demographics: • Age • Gender • Area of residence B.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
131
+ page_content=' Getting information routine wise online learning during the shift from face to face study to online study in colleges/Universities in Pakistan.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
132
+ page_content=' These information consists of the following: • Average time given for online study in hours per day • Quality and the problems of the communication medium • Actual involvement in virtual lecture same as face to face lecture in physical class • Level of interruption by the family members during online study period • Attention and focus level from joining to the end of online class.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
133
+ page_content=' • Effects of online learning on Cheating behavior and students involvement to C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
134
+ page_content=' Evaluation of the experience of the student’s level of involvement in virtual class to find the overall students involvement in online lecture.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
135
+ page_content=' D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
136
+ page_content=' Evaluation of health during change in learning style from physical class environment provided by the College/University to the virtual class environment provided by your parents at home and the effects of virtual class on your involvement of class.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
137
+ page_content=' ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII WEB: https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
138
+ page_content='ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
139
+ page_content='org E-MAIL: bilgi@ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
140
+ page_content='org 1661 The pictorial survey responses are given below: Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
141
+ page_content=' 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
142
+ page_content=' Getting General Info Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
143
+ page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
144
+ page_content=' Getting General Info Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
145
+ page_content=' 3(a).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
146
+ page_content=' Getting Specific Info Section A: Demographics Info Chart 100% 80% 60% 40% 20% 0% 6 1 4 4 5 8 5 9 3 2 3 3 3 4 4 4 5 5 5 Gender Age Degree_ Level Area of ResidenceSection B: Getting General Info Chart 100% 80% 60% 40% 20% 0% 3 5 9 2 8 5 8 385 417 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
147
+ page_content='49 3 4 8 3 6 1 1 T 2 3 3 5 5 5 6 Time_Spent_SociaMediaLaptopComputerAvail ISmartPhonesAvail Class Participation Level CheatingConcern StudyLevelifdontExam I Lack of IT Skills BetterOnlineLearnSectionC:GettingSpecificInfo(11-17 Questions) 100% 80% 60% 40% 20% 0% 8 6 OnlineAndOffineEqual Technicallssuelmpact Economiclssuelmpact TeacherVoicelssue Lessrsinteraction AcademicLossfearinClassParticipation LackDeficiencyforNonITSt ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
148
+ page_content='ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII WEB: https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
149
+ page_content='ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
150
+ page_content='org E-MAIL: bilgi@ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
151
+ page_content='org 1662 We have created questionnaire.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
152
+ page_content=' Its soft copy is available at the following link: https://docs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
153
+ page_content='google.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
154
+ page_content='com/forms/d/1zqnXC9EXRXjmNL7VX2FP4hh6OL0NVu- C_w8QZOFxRsc/edit ) We have collected 623 responses from different level of degree students.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
155
+ page_content=' Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
156
+ page_content=' 3(b).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
157
+ page_content=' Getting Specific Info Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
158
+ page_content=' 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
159
+ page_content=' Getting Health Issues info V.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
160
+ page_content=' EXPEIRMENTAL RESULTS The means and standard deviation of all the variables as per questionnaire are given below: ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
161
+ page_content='SectionC:GettingSpecificInfo(1-10Questions) 100% 80% 60% 40% 20% 0% 8 5 2 5 BetterTimeUtilization CheatingBehavior Umwilingness_of_ResponsibilityStudentsHesitancylmpact TechDificultylmpact HaveNetAccess HaveElectricSuppy InteractionWihTeacher ClassParticipationChance AttensionAndFocusDisturbSection D: Getting Health Issues Info Chart 100% 80% 60% 40% 20% 0% 4 8 2 3 80 365 393 2 4 3 89 9 3 3 4 5 5 5 5 6 FitnessOfAttensionLonelinessEffectAnxietyLevelPsycholmpacts ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
162
+ page_content='ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII WEB: https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
163
+ page_content='ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
164
+ page_content='org E-MAIL: bilgi@ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
165
+ page_content='org 1663 TABLE I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
166
+ page_content=' S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
167
+ page_content='# Variable Value Mean of all the variables SECTION A: DEMOGRAPHICS INFO 1 Gender 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
168
+ page_content='400000 2 Age 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
169
+ page_content='028571 3 Degree_Level 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
170
+ page_content='257143 4 Area_of_Residence 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
171
+ page_content='628571 SECTION B: GETTING GENERAL INFO 1 Time_Spent_ SociaMedia 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
172
+ page_content='457143 2 LaptopComputerAvail 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
173
+ page_content='771429 3 SmartPhonesAvail 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
174
+ page_content='571429 4 Class_Participation_Level 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
175
+ page_content='885714 5 CheatingConcern 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
176
+ page_content='942857 6 StudyLevelIfdontExam 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
177
+ page_content='028571 7 Lack_of_IT_Skills 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
178
+ page_content='485714 8 BetterOnlineLearn 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
179
+ page_content='600000 SECTION C: GETTING SPECIFIC INFO 1 BetterTimeUtilization 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
180
+ page_content='342857 2 CheatingBehavior 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
181
+ page_content='285714 3 Unwilingness_of_Responsibility 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
182
+ page_content='114286 4 StudentsHesitancyImpact 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
183
+ page_content='371429 5 TechDifficultyImpact 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
184
+ page_content='200000 6 HaveNetAccess 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
185
+ page_content='514286 7 HaveElectricSupply 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
186
+ page_content='000000 8 InteractionWihTeacher 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
187
+ page_content='085714 9 ClassParticipationChance 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
188
+ page_content='057143 10 AttensionAndFocusDisturb 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
189
+ page_content='342857 11 OnlineAndOfflineEqual 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
190
+ page_content='800000 12 TechnicalIssueImpact 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
191
+ page_content='857143 13 EconomicIssueImpact 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
192
+ page_content='000000 14 TeacherVoiceIssue 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
193
+ page_content='971429 15 LessTSInteraction 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
194
+ page_content='857143 16 AcademicLossFearinClassParticipation 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
195
+ page_content='028571 16 AcademicLossFearinClassParticipation 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
196
+ page_content='970588 17 LackDeficiencyforNonITSt 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
197
+ page_content='747240 Standard Deviation of all the variables SECTION A: DEMOGRAPHICS INFO 1 Gender 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
198
+ page_content='489898 2 Age 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
199
+ page_content='376883 3 Degree_Level 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
200
+ page_content='552545 4 Area_of_Residence 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
201
+ page_content='483187 ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII WEB: https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
202
+ page_content='ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
203
+ page_content='org E-MAIL: bilgi@ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
204
+ page_content='org 1664 SECTION B: GETTING GENERAL INFO 1 Time_Spent_ SociaMedia 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
205
+ page_content='078169 2 LaptopComputerAvail 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
206
+ page_content='897161 3 SmartPhonesAvail 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
207
+ page_content='766652 4 Class_Participation_Level 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
208
+ page_content='259738 5 CheatingConcern 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
209
+ page_content='093954 6 StudyLevelIfdontExam 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
210
+ page_content='502107 7 Lack_of_IT_Skills 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
211
+ page_content='105090 8 BetterOnlineLearn 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
212
+ page_content='515633 SECTION C: GETTING SPECIFIC INFO 1 BetterTimeUtilization 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
213
+ page_content='413059 2 CheatingBehavior 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
214
+ page_content='110249 3 Unwilingness_of_Responsibility 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
215
+ page_content='259738 4 StudentsHesitancyImpact 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
216
+ page_content='332789 5 TechDifficultyImpact 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
217
+ page_content='979796 6 HaveNetAccess 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
218
+ page_content='273273 7 HaveElectricSupply 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
219
+ page_content='309307 8 InteractionWihTeacher 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
220
+ page_content='295518 9 ClassParticipationChance 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
221
+ page_content='286032 10 AttensionAndFocusDisturb 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
222
+ page_content='392692 11 OnlineAndOfflineEqual 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
223
+ page_content='214202 12 TechnicalIssueImpact 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
224
+ page_content='797957 13 EconomicIssueImpact 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
225
+ page_content='956183 3 TeacherVoiceIssue 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
226
+ page_content='027777 4 LessTSInteraction 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
227
+ page_content='045886 TABLE II: TTest Outcomes Ttest_1sampResult(statistic=array([16.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
228
+ page_content='663333 , 31.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
229
+ page_content='38507589, 23.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
230
+ page_content='81939622, 19.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
231
+ page_content='65311057, 13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
232
+ page_content='28871279, 11.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
233
+ page_content='51311097, 11.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
234
+ page_content='95187108, 13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
235
+ page_content='35711613, 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
236
+ page_content='35574591, 11.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
237
+ page_content='7564528 , 13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
238
+ page_content='11574349, 13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
239
+ page_content='84994208, 13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
240
+ page_content='79421828, 12.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
241
+ page_content='0044142 , 9.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
242
+ page_content='78640192, 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
243
+ page_content='375 , 13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
244
+ page_content='09261879, 11.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
245
+ page_content='51416659, 13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
246
+ page_content='36038922, 13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
247
+ page_content='88838218, 13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
248
+ page_content='86128572, 9.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
249
+ page_content='80912102, 18.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
250
+ page_content='24871239, 13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
251
+ page_content='57080199, 12.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
252
+ page_content='19631092, 11.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
253
+ page_content='18462458, 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
254
+ page_content='35381536, 12.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
255
+ page_content='18694645, 13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
256
+ page_content='15416906, 11.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
257
+ page_content='9272551 , 11.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
258
+ page_content='34226868, 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
259
+ page_content='64348064, 11.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
260
+ page_content='2720409 ]), pvalue=array([6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
261
+ page_content='29551067e- 18, 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
262
+ page_content='08636299e-26, 8.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
263
+ page_content='60469002e-23, 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
264
+ page_content='85366635e-20, 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
265
+ page_content='07753770e-15, 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
266
+ page_content='80770597e-13, 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
267
+ page_content='00580609e-13, 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
268
+ page_content='38220401e-15, 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
269
+ page_content='72979440e-12, 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
270
+ page_content='58421613e-13, 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
271
+ page_content='38588068e-15, 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
272
+ page_content='53985258e-15, 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
273
+ page_content='73086219e-15, 8.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
274
+ page_content='90861219e-14, 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
275
+ page_content='02190243e-11, 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
276
+ page_content='50637008e-12, 7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
277
+ page_content='76732461e-15, 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
278
+ page_content='80069994e-13, 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
279
+ page_content='35148773e-15, 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
280
+ page_content='42079500e-15, 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
281
+ page_content='50369222e-15, 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
282
+ page_content='90647656e-11, 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
283
+ page_content='87615125e-19, 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
284
+ page_content='77545554e-15, 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
285
+ page_content='73530395e-14, 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
286
+ page_content='15085107e-13, 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
287
+ page_content='75281129e-12, 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
288
+ page_content='85928814e-14, 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
289
+ page_content='79392859e-15, 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
290
+ page_content='06477004e-13, 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
291
+ page_content='21455437e-13, 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
292
+ page_content='30654437e-12, 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
293
+ page_content='98568395e-13])) ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII WEB: https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
294
+ page_content='ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
295
+ page_content='org E-MAIL: bilgi@ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
296
+ page_content='org 1665 VIII.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
297
+ page_content=' CONCLUSTION To evaluate and find the correctness and applicability of the hypothesis as per the problem statement, we used an online survey approach using Google docs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
298
+ page_content=' According to the percentages of survey responses given in Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
299
+ page_content='1 Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
300
+ page_content='2, Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
301
+ page_content='3(a), Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
302
+ page_content='3(b) and Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
303
+ page_content=' 4, availability of Laptop/Computer at student homes was 85.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
304
+ page_content='3% and smart phones was 87.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
305
+ page_content='5%.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
306
+ page_content=' Time spent on social media during the online lecture was 71%.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
307
+ page_content=' The level of Class participation was 49.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
308
+ page_content='6%.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
309
+ page_content=' The concern of students cheating during the online exam was 70.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
310
+ page_content='8%—level of cheating behavior to ignore interest in online due to online exams encouraged by 60.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
311
+ page_content='8% of students.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
312
+ page_content=" The student's unwillingness was found at 73." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
313
+ page_content='2%.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
314
+ page_content=' Impacts of technical issues during online classes were 84.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
315
+ page_content='4% .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
316
+ page_content=" The pace of the teacher's voice due to the Net problem was discovered at 79." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
317
+ page_content='5% and Impacts of less interaction of teacher-student found to be 78%.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
318
+ page_content=' As per Fig.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
319
+ page_content=' 4, psychological impacts on learning participation during online classes were discovered at 73.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
320
+ page_content="5% , the stress of loneliness affects students' level of involvement was 68." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
321
+ page_content="5% and the anxiety levels disturb students' level of motivation by 77." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
322
+ page_content='9%.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
323
+ page_content=' As per the above experiments, the means and standard deviations are given in Table I and Table II above.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
324
+ page_content=' Most of the high values of means shows that much percentage of the students are not fully involved during online lecture.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
325
+ page_content=' Similarly, most of the values of standard deviations are far from zero.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
326
+ page_content=' It shows that data points are far from the mean.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
327
+ page_content=" We applied the test here because we don't have actual population means." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
328
+ page_content=" We used ttest_1samp (Dataset [:35],0) with a default value of 0 to compute the variables' statistics and p-value." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
329
+ page_content=' The results of this test are provided in Table II.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
330
+ page_content=' These values are minimal, which is in favor of rejecting the null or H0 hypothesis and accepting the alternate or H1 hypothesis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
331
+ page_content=" It further means that students' involvement during online classes is severely affected." metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
332
+ page_content=' ACKNOWLEDGMENT The authors are very grateful to the management of server room of Faculty of Computing and Information Technology (FCIT), University of the Punjab to forward our questionnaire to the students for the responses.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
333
+ page_content=' We are also thankful to the Students of Undergraduates and Gradates students of FCIT for the warm participation and sincere responses during the survey of this research study.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
334
+ page_content=' REFERENCES [1].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
335
+ page_content=' Fernando, R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
336
+ page_content=', “The COVID-19 Pandemic: A call for a reality check.”, published in Galle Medical Journal, 25 (1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
337
+ page_content=' 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
338
+ page_content=' [2].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
339
+ page_content=' Myers, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
340
+ page_content=', “After COVID-19: Recalibrating the American educational system”, Retrieved from https://hub.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
341
+ page_content='jhu.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
342
+ page_content='edu/2020/04/07/bob-balfanz-education-reform-covid- 19/, accessed 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
343
+ page_content=' [3].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
344
+ page_content=' Tam, G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
345
+ page_content=', & El-Azar, D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
346
+ page_content=' (2020), “3 ways the coronavirus pandemic could reshape education”, Retrieved from https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
347
+ page_content='weforum.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
348
+ page_content='org/agenda/2021/05/consumer- demand-covid-19-recovery/ published in 2020 but accessed 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
349
+ page_content=' [4].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
350
+ page_content=' World Bank, “How countries are using edtech (including online learning, radio, television, texting) to support access to remote learning during the COVID-19 pandemic.”, Retrieved from https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
351
+ page_content='worldbank.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
352
+ page_content='org/en/topic/edutech/brief/how- countries-are-using-edtech-to-support-remote-learning-during-the-covid-19-pandemic , published in 2020 but accessed 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
353
+ page_content=' [5].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
354
+ page_content=' Ractham, P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
355
+ page_content=', & Chen, C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
356
+ page_content=', “Promoting the use of online social technology as a case-based learning tool”, published in Journal of Information Systems Education, 24 (4), 2019.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
357
+ page_content=' ANKARA INTERNATIONAL CONGRESS ON SCIENTIFIC RESEARCH-VII WEB: https://www.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
358
+ page_content='ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
359
+ page_content='org E-MAIL: bilgi@ankarakongresi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
360
+ page_content='org 1666 [6].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
361
+ page_content=' Noor Ul Ain, “Is online education the new future of Pakistan?”' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
362
+ page_content=' Retrieved from https://dailytimes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
363
+ page_content='com.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
364
+ page_content='pk/579663/blessing-in-disguise-is-online-education-the-new-future-of- pakistan/ , published in 2020 but accessed 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
365
+ page_content=' [7].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
366
+ page_content=' Noor Ul Ain Ali.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
367
+ page_content=', “Students disappointed with online teaching system amid COVID-19”.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
368
+ page_content=' Retrieved from https://dailytimes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
369
+ page_content='com.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
370
+ page_content='pk/587446/students-disappointed-with- online-teachingsystem-amid-covid-19/, published in 2020 but accessed 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
371
+ page_content=' [8].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
372
+ page_content=' Kebritchi, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
373
+ page_content=', Lipschuetz, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
374
+ page_content=', & Santiague, L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
375
+ page_content=', “Issues and challenges for teaching successful online courses in higher education: A literature review”, Journal of Educational Technology Systems, 46 (1), 4–29, 2017.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
376
+ page_content=' [9].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
377
+ page_content=' Kunal Chaturvedi, Dinesh Kumar Vishwakarma, Nidhi Singh, “ COVID-19 and its impact on education, social life and mental health of students: A survey”, published in Children and Youth Services Review 121 (2021) 105866, pp.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
378
+ page_content=' 1-6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
379
+ page_content=' [10].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
380
+ page_content=' Nico W.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
381
+ page_content=' Van Yperen, Melvyn R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
382
+ page_content='W.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
383
+ page_content=' Hamstra and Marloes van der Klauw, “ To Win, or Not to Lose, At Any Cost: The Impact of Achievement Goals on Cheating”, published in British Journal of Management, Vol.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
384
+ page_content=' 22, S5–S15 (2011) [11].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
385
+ page_content=' Lia M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
386
+ page_content=' Daniels, Lauren D, Goeganx Patti C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
387
+ page_content=' Parker, “ The impact of COVID 19 triggered changes to instruction and assessment on university students’ self reported motivation, engagement and perceptions”, published in Social Psychology of Education (2021) 24:299–318, Springer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
388
+ page_content=' [12].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
389
+ page_content=' Timon ElmerID, Kieran Mepham, Christoph Stadtfeld, “Students under lockdown: Comparisons of students’ social networks and mental health before and during the COVID-19 crisis in Switzerland”, published in PLOS ONE, pp: 1-22, 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
390
+ page_content=' [13].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
391
+ page_content=' Mohammed Amin Almaiah, Ahmad Al-Khasawneh and Ahmad Althunibat, “ Exploring the critical challenges and factors influencing the E-learning system usage during COVID-19 pandemic”, published in Journal of Education and Information Technologies, Springer.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
392
+ page_content=' [14].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
393
+ page_content=' Huma Sarwar, Hira Akhtar, Meshal Muhammad Naeem, Javeria Ali Khan, Khadija Waraich, Sumaiya Shabbir, Arshad Hasan and Zohaib Khurshid, “ Self-Reported Effectiveness of e-Learning Classes during COVID-19 Pandemic: A Nation-Wide Survey of Pakistani Undergraduate Dentistry Students”, published in European Journal of Dentistry, 2020;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
394
+ page_content='14(suppl S1):S34–S43.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
395
+ page_content=' [15].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
396
+ page_content=' Aqsa Arshad, Madiha Afzal, * Dr.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
397
+ page_content=' Muhammad Sabboor Hussain, “ Sudden Switch to Post-COVID-19 Online Classes and Cognitive Transformation of ESL Learners: Critical Analysis of Discourse of Fear”, published in Research Journal of Social Sciences & Economics Review, Vol.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
398
+ page_content=' 1, Issue 3, 2020, PP: 188-199.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
399
+ page_content=' [16].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
400
+ page_content=' Najmul Hasan, Yukun Bao, “ Impact of “e-Learning crack-up” perception on psychological distress among college students during COVID-19 pandemic: A mediating role of “fear of academic year loss”, published in Children and Youth Services Review 118 (2020) PP: 1-10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
401
+ page_content=' [17].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
402
+ page_content=' Muhammad Anwar, Anwar Khan, Khalid Sultan, “The barriers and challenges faced by students in online education during covid-19 pandemic in Pakistan”, published in Gomal University Journal of Research, Volume 36, Issue 1, JUNE, 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
403
+ page_content=' PP.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
404
+ page_content=' 52-62.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
405
+ page_content=' [18].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
406
+ page_content=' Brochure, what is a survey?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
407
+ page_content=' Bill Kalsbeek, 1995 publications officer, ASA section on survey research methods, 1995.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/2tAyT4oBgHgl3EQfPvai/content/2301.00031v1.pdf'}
3dE3T4oBgHgl3EQfPwmd/content/tmp_files/2301.04406v1.pdf.txt ADDED
@@ -0,0 +1,794 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ arXiv:2301.04406v1 [cs.DS] 11 Jan 2023
2
+ A Note on Property Testing of the Binary Rank
3
+ Nader H. Bshouty
4
+ Dept. of Computer Science
5
+ Technion, Haifa, Israel.
6
+ January 12, 2023
7
+ Abstract
8
+ Let M be a n × m (0, 1)-matrix. We define the s-binary rank, brs(M), of M to be the
9
+ minimal integer d such that there are d monochromatic rectangles that cover all the 1-entries in
10
+ the matrix, and each 1-entry is covered by at most s rectangles. When s = 1, this is the binary
11
+ rank, br(M), known from the literature.
12
+ Let R(M) and C(M) be the set of rows and columns of M, respectively. We use the result of
13
+ Sgall [8] to prove that if M has s-binary rank at most d, then |R(M)| · |C(M)| ≤
14
+ � d
15
+ ≤s
16
+
17
+ 2d where
18
+ � d
19
+ ≤s
20
+
21
+ = �s
22
+ i=0
23
+ �d
24
+ i
25
+
26
+ . This bound is tight; that is, there exists a matrix M ′ of s-binary rank d such
27
+ that |R(M ′)| · |C(M ′)| =
28
+ � d
29
+ ≤s
30
+
31
+ 2d.
32
+ Using this result, we give a new one-sided adaptive and non-adaptive testers for (0, 1)-
33
+ matrices of s-binary rank at most d (and exactly d) that makes ˜O
34
+ �� d
35
+ ≤s
36
+
37
+ 2d/ǫ
38
+
39
+ and ˜O
40
+ �� d
41
+ ≤s
42
+
43
+ 2d/ǫ2�
44
+ queries, respectively.
45
+ For a fixed s, this improves the query complexity of the tester of Parnas et al. in [7] by a
46
+ factor of ˜Θ(2d).
47
+ 1
48
+ Introduction
49
+ Let M be a n × m (0, 1)-matrix. We define the s-binary rank, brs(M), of M to be the minimal
50
+ integer d such that there are d sets (rectangles) Ik×Jk where Ik ⊆ [n] := {1, . . . , n}, Jk ⊂ [m], k ∈ [d]
51
+ such that1 M[i, j] = 1 for all (i, j) ∈ Ik × Jk, k ∈ [d] (monochromatic rectangles), and for every
52
+ (i, j) ∈ [n] × [m] where M[i, j] = 1, there are at least one and at most s integers t ∈ [d] such
53
+ that (i, j) ∈ It × Jt (each 1-entry in M is covered by at least one and at most s monochromatic
54
+ rectangles). When s = 1, br1(M), is the binary rank, br(M), and when s = ∞, br∞(M) is the
55
+ Boolean rank. Both are known from the literature. See, for example, [4].
56
+ The binary rank can also be defined as follows. The binary rank of a n × m (0, 1)-matrix M
57
+ is equal to the minimal d, where there are n × d (0, 1)-matrix N and d × m (0, 1)-matrix L such
58
+ that M = NL. It is also equal to the minimal number of bipartite cliques needed to partition all
59
+ the edges of a bipartite graph whose adjacent matrix is M. The s-binary rank of M is the minimal
60
+ number of bipartite cliques needed to cover all edges of a bipartite graph whose adjacent matrix
61
+ is M, where each edge is covered by at most s bipartite cliques. In [2], it was shown that it is
62
+ NP-hard to approximating the binary rank to within a factor of n1−δ for any given δ.
63
+ 1For M, the (i, j) entry of the matrix is denoted by M[i, j].
64
+ 1
65
+
66
+ A property-testing algorithm (tester) of the s-binary rank [7] is given as input 0 < ǫ < 1, integers
67
+ d, n, m, and query access to the entries of a n×m (0, 1)-matrix M. If M has s-binary rank at most
68
+ d (resp. equal d), then the tester accepts with probability at least 2/3. If M is ǫ-far from having
69
+ s-binary rank at most d (resp. equal d), i.e., more than ǫ-fraction of the entries of M should be
70
+ modified to get a matrix with s-binary rank at most d (resp. equal to d), then the tester rejects
71
+ with probability at least 2/3. If the tester accepts matrices having s-binary rank at most d (resp.
72
+ equal to d) with probability 1, then we call it a one-sided error tester. In adaptive testing, the
73
+ queries can depend on the answers to the previous queries, whereas in non-adaptive testing, all the
74
+ queries are fixed in advance by the tester. The goal is to construct a tester that makes a minimal
75
+ number of queries.
76
+ The testability of s-binary rank at most d of (0, 1)-matrices was studied in [6, 7]. In [6], Nakar
77
+ and Ron gave a non-adaptive one-sided error tester for s = 1, that makes ˜O(24d/ǫ4). In [7], Parnas
78
+ et al. gave a non-adaptive and adaptive one-sided error tester for s = 1 that makes O(22d/ǫ2) and
79
+ O(22d/ǫ) queries, respectively. The results in [7] also hold for s-binary rank at most d. In this
80
+ paper, for s-binary at most d and equal to d, we prove
81
+ Theorem 1. There exists an adaptive one-sided error tester for s-binary rank of n × m (0, 1)-
82
+ matrices that makes ˜O
83
+ �� d
84
+ ≤s
85
+
86
+ 2d/ǫ
87
+
88
+ queries.
89
+ Theorem 2. There exists a non-adaptive one-sided error tester for s-binary rank of n × m (0, 1)-
90
+ matrices that makes ˜O
91
+ �� d
92
+ ≤s
93
+
94
+ 2d/ǫ2�
95
+ queries.
96
+ For fixed s, this improves the query complexity of Parnas et al. in [7] by a factor of ˜O(2d).
97
+ 1.1
98
+ Our Approach
99
+ The tester of Parnas et al. [7] uses the fact that if M′ is a k × k sub-matrix of M and M′ is of
100
+ s-binary rank at most d, then
101
+ 1. M′ has at most 2d distinct rows and at most 2d distinct columns.
102
+ 2. If M is ǫ-far from having s-binary rank at most d, then extending M′ by one more uniformly
103
+ at random row and column of M, gives a (k + 1) × (k + 1) sub-matrix M′′ of M that, with
104
+ probability at least Ω(ǫ), satisfies: the number of distinct rows in M′′ is greater by one than
105
+ the number of distinct rows in M′, or, the number of distinct columns in M′′ is greater by
106
+ one than the number of distinct columns in M′.
107
+ So, their adaptive tester runs O(2d/ǫ) iterations. At each iteration, it extends M′ by uniformly at
108
+ random one row and one column. Let M′′ be the resulting sub-matrix. If the s-binary rank of M′′
109
+ is greater than d, the tester rejects. If the number of distinct rows or columns in M′′ is greater
110
+ than the number in M′, then it continues to the next iteration with M′ ← M′′. Otherwise, it
111
+ continues to the next iteration with M′. If, after O(2d/ǫ) iterations, M′ has s-binary rank d, the
112
+ tester accepts.
113
+ If the s-binary rank of M is d, then every sub-matrix has a s-binary rank d, and the tester
114
+ accepts. If M is ǫ-far from having s-binary rank at most d, then: since, at each iteration, with
115
+ probability at least Ω(ǫ), the number of distinct rows or columns of M′ is increased by one, and
116
+ since matrices of s-binary rank d has at most 2d distinct rows and at most 2d distinct columns,
117
+ with high probability, we get M′ with s-binary rank greater than d and the tester rejects. The
118
+ 2
119
+
120
+ query complexity of the tester is O(22d/ǫ), which is the number of entries of the matrix M′, O(22d),
121
+ times the number of trials O(1/ǫ) for extending M′ by one row and one column.
122
+ We now give our approach. Call a sub-matrix M′ of M perfect if it has distinct rows and distinct
123
+ columns. Our adaptive tester uses the fact that if M′ is a perfect k×k′ sub-matrix of M of s-binary
124
+ rank d, then
125
+ 1. kk′ ≤
126
+ � d
127
+ ≤s
128
+
129
+ 2d.
130
+ 2. If M is ǫ-far from having s-binary rank at most d, then at least one of the following occurs
131
+ (a) With probability at least Ω(ǫ), extending M′ by one uniformly at random column of M,
132
+ gives a perfect k × (k′ + 1) sub-matrix M′′ of M.
133
+ (b) With probability at least Ω(ǫ), extending M′ by one uniformly at random row of M,
134
+ gives a perfect (k + 1) × k′ sub-matrix M′′ of M.
135
+ (c) With probability at least Ω(ǫ), extending M′ by one uniformly at random column and
136
+ one uniformly at random row of M, gives a perfect2 (k + 1) × (k′ + 1) sub-matrix M′′ of
137
+ M.
138
+ Item 1 follows from Sgall result in [8] (See Section 3), and item 2 is Claim 10 in [7]. Now, the
139
+ tester strategy is as follows. If k ≤ k′, the tester first tries to extend M′ with a new column. If
140
+ it succeeds, it moves to the next iteration. Otherwise, it tries to extend M′ with a new row. If it
141
+ succeeds, it moves to the next iteration. Otherwise, it tries to extend M′ with a new row and a
142
+ new column. If it succeeds, it moves to the next iteration. If it fails, it accepts. If k′ < k, it starts
143
+ with the row, then the column, and then both.
144
+ Using this strategy, we show that the query complexity will be, at most, the order of the size
145
+ kk′ ≤
146
+ � d
147
+ ≤s
148
+
149
+ 2d of M′ times the number of trials, ˜O(1/ǫ), to find the new row, column, or both. This
150
+ achieves the query complexity in Theorem 1.
151
+ For the non-adaptive tester, the tester, uniformly at random, chooses t = ˜O
152
+ �� d
153
+ ≤s
154
+
155
+ 2d/ǫ2�
156
+ rows
157
+ r1, . . . , rt ∈ [n] and t columns c1, . . . , ct ∈ [m] and queries all M[ri, cj] for all i · j ≤ t and puts them
158
+ in a table. Then it runs the above non-adaptive tester. When the non-adaptive tester asks for
159
+ uniformly at random row or column, it provides the next element ri or cj, respectively. The queries
160
+ are then answered from the table. We show that the adaptive algorithm does not need to make
161
+ queries that are not in the table before it halts. This achieves the query complexity in Theorem 2.
162
+ 1.2
163
+ Other Rank Problems
164
+ The real rank of a n × m-matrix M over any field F is the minimal d, such that there is a n × d
165
+ matrix N over F and a d × m matrix L over F such that M = NL. The testability of the real
166
+ rank was studied in [1, ?, 5]. In [1], Balcan et al. gave a non-adaptive tester for the real rank that
167
+ makes ˜O(d2/ǫ) queries. They also show that this query complexity is optimal.
168
+ The Boolean rank (∞-binary rank) was studied in [6, 7]. Parnas et al. in [7] gave a non-adaptive
169
+ tester for the Boolean rank that makes ˜O(d4/ǫ4) queries3.
170
+ 2It may happen that events (a) and (b) do not occur and (c) does
171
+ 3The query complexity in [7] is ˜O(d4/ǫ6).
172
+ We’ve noticed that Lemma 3 in [7] is also true when we replace
173
+ (ǫ2/64)n2 with (ǫ/4)n2. To prove that, in the proof of Lemma 3, replace Modification rules 1 and 2 with the following
174
+ modification: Modify to 0 all beneficial entries. This gives the result stated here,[3].
175
+ 3
176
+
177
+ 2
178
+ Definitions and Preliminary Results
179
+ Let M be a n × m (0, 1)-matrix. We denote by R(M) and C(M) the set of rows and columns
180
+ of M, respectively. The number of distinct rows and columns of M are denoted by r(M) = |R(M)|
181
+ and, c(M) = |C(M)|, respectively. The binary rank of a n × m-matrix M, br(M), is equal to the
182
+ minimal d, where there is a n × d (0, 1)-matrix N and a d × m (0, 1)-matrix L such that M = NL.
183
+ We define the s-binary rank, brs(M), of M to be the minimal integer d such that there are d
184
+ sets (rectangles) Ik × Jk where Ik ⊆ [n] := {1, . . . , n}, Jk ⊂ [m], k ∈ [d] such that M[i, j] = 1 for all
185
+ (i, j) ∈ Ik ×Jk, k ∈ [d] (monochromatic rectangles) and for every (i, j) ∈ [n]×[m] where M[i, j] = 1
186
+ there are at least one and at most s integers t ∈ [d] such that (i, j) ∈ It × Jt (each 1-entry in M is
187
+ covered by at least one and at most s monochromatic rectangles).
188
+ We now prove.
189
+ Lemma 1. Let M be a n × m (0, 1)-matrix. The s-binary rank of M, brs(M), is equal to the
190
+ minimal integer d, where there is a n × d (0, 1)-matrix N and a d × m (0, 1)-matrix L such that:
191
+ For P = NL,
192
+ 1. For every (i, j) ∈ [n] × [m], M[i, j] = 0 if and only if P[i, j] = 0.
193
+ 2. For every (i, j) ∈ [n] × [m], P[i, j] ≤ s.
194
+ Proof. If M is of s-binary rank d, then there are rectangles {Ik ×Jk}k∈[d], Ik ⊆ [n], Jk ⊂ [m], k ∈ [d]
195
+ such that M[i, j] = 1 for all (i, j) ∈ Ik ×Jk, k ∈ [d] and for every (i, j) ∈ [n]×[m] where M[i, j] = 1
196
+ there are at least one and at most s integers t ∈ [d] such that (i, j) ∈ It × Jt. Define row vectors
197
+ a(k) ∈ {0, 1}n and b(k) ∈ {0, 1}m where a(k)
198
+ i
199
+ = 1 iff (if and only if) i ∈ Ik, and b(k)
200
+ j
201
+ = 1 iff j ∈ Jk.
202
+ Then define4 P = a(1)′b(1)+· · ·+a(d)′b(d). It is easy to see that (a(k)′b(k))[i, j] = 1 iff (i, j) ∈ Ik ×Jk.
203
+ Therefore, P[i, j] = 0 iff M[i, j] = 0 and P[i, j] ≤ s for all (i, j) ∈ [n]×[m]. Define the n×d matrix
204
+ N =
205
+
206
+ a(1)′| · · · |a(d)′�
207
+ and the d × m matrix L =
208
+
209
+ b(1)′| · · · |b(d)′�′
210
+ .
211
+ It is again easy to see that
212
+ P = NL.
213
+ The other direction can be easily seen by tracing backward in the above proof.
214
+ We now prove the following,
215
+ Lemma 2. Let M be a n × m matrix. Let N and L be n × d (0, 1)-matrix and d × m (0, 1)-matrix,
216
+ respectively, such that P = NL. Then r(P) ≤ r(N) and c(P) ≤ c(L).
217
+ Proof. We prove the result for r.
218
+ The proof for c is similar.
219
+ Let r1, . . . , rn be the rows of N
220
+ and p1, . . . , pn be the rows of P.
221
+ Then pi = riL.
222
+ Therefore, if ri = rj, then pi = pj.
223
+ Thus,
224
+ r(P) ≤ r(N).
225
+ Let M be a n × m matrix. For x ∈ X ⊆ [n], y ∈ Y ⊆ [m], we denote by M[X, Y ] the |X| × |Y |
226
+ sub-matrix of M, (M[x′, y′])x′∈X,y′∈Y . Denote by M[X, y] the column vector (M[x′, y])x′∈X and by
227
+ M[x, Y ] the row vector (M(x, y′))y′∈Y .
228
+ For x ∈ [n] (resp. y ∈ [m]) we say that M[X, y] is a new column (resp. M[x, Y ] is a new row)
229
+ to M[X, Y ] if it is not equal to any of the columns (resp. rows) of M[X, Y ].
230
+ 4Here x′ is the transpose of x.
231
+ 4
232
+
233
+ Lemma 3. Let M be a n × m matrix, x ∈ [n], X ⊆ [n], y ∈ [m], and Y ⊆ [m]. Suppose M[x, Y ] is
234
+ not a new row to M[X, Y ], and M[X, y] is not a new column to M[X, Y ]. Then M[x, Y ∪ {y}] is
235
+ not a new row to M[X, Y ∪{y}] if and only if M[X ∪{x}, y] is not a new column to M[X ∪{x}, Y ].
236
+ Proof. If M[x, Y ∪ {y}] is not a new row to M[X, Y ∪ {y}], then there is x′ ∈ X such that
237
+ M[x, Y ∪ {y}] = M[x′, Y ∪ {y}]. Since M[X, y] is not a new column to M[X, Y ], there is y′ ∈ Y
238
+ such that M[X, y] = M[X, y′]. Since M[x, Y ∪ {y}] = M[x′, Y ∪ {y}], we have M[x′, y′] = M[x, y′]
239
+ and M[x, y] = M[x′, y].
240
+ Since M[X, y] = M[X, y′], we have M[x′, y] = M[x′, y′].
241
+ Therefore,
242
+ M[x, y] = M[x, y′] and M[X ∪ {x}, y] = M[X ∪ {x}, y′]. Thus, M[X ∪ {x}, y] is not a new column
243
+ to M[X ∪ {x}, Y ].
244
+ Similarly, the other direction follows.
245
+ 3
246
+ Matrices of s-Binary Rank d
247
+ In this section, we prove the following two Lemmas.
248
+ Lemma 4. For any n × m (0, 1)-matrix M of s-binary rank at most d, we have
249
+ r(M) · c(M) ≤
250
+ � d
251
+ ≤ s
252
+
253
+ 2d.
254
+ Lemma 5. There is a (0, 1)-matrix M′ of s-binary rank d that satisfies r(M′) · c(M′) =
255
+ � d
256
+ ≤s
257
+
258
+ 2d.
259
+ To prove Lemma 4, we use the following Sgall’s lemma.
260
+ Lemma 6. [8]. Let A, B ⊆ 2[d] be such that for every A ∈ A and B ∈ B, |A ∩ B| ≤ s. Then
261
+ |A| · |B| ≤
262
+ � d
263
+ ≤s
264
+
265
+ 2d.
266
+ We now prove Lemma 4.
267
+ Proof. Since the s-binary rank of M is at most d, by Lemma 1, there is a n × d (0, 1)-matrix N
268
+ and a d × m (0, 1)-matrix L such that, for P = NL
269
+ 1. For every (i, j) ∈ [n] × [m], M[i, j] = 0 if and only if P[i, j] = 0.
270
+ 2. For every (i, j) ∈ [n] × [m], P[i, j] ≤ s.
271
+ Obviously, r(M) ≤ r(P) and c(M) ≤ c(P).
272
+ Consider A = {A1, . . . , An} ⊆ 2[d] and B =
273
+ {B1, . . . , Bm} ⊆ 2[d], where Ai = {j|Ni,j = 1} and Bk = {j|Lj,k = 1}.
274
+ Since the entries of
275
+ P = NL are at most s, for every i ∈ [n] and k ∈ [m], |Ai ∩ Bk| ≤ s.
276
+ By Lemma 2 and 6,
277
+ r(M) · c(M) ≤ r(P) · c(P) ≤ r(N) · c(L) = |A| · |B| ≤
278
+ � d
279
+ ≤ s
280
+
281
+ 2d.
282
+ We now prove Lemma 5
283
+ 5
284
+
285
+ Proof. Let N be a 2d × d (0, 1)-matrix where its rows contain all the vectors in {0, 1}d. Let L be a
286
+ d ×
287
+ � d
288
+ ≤s
289
+
290
+ matrix where its columns contain all the vectors in {0, 1}d of weight at most s. Obviously,
291
+ P = NL is 2d ×
292
+ � d
293
+ ≤s
294
+
295
+ with entries that are less than or equal to s. Define a 2d ×
296
+ � d
297
+ ≤s
298
+
299
+ (0, 1)-matrix
300
+ M′ where M′[i, j] = 0 if and only if P[i, j] = 0. Then, by Lemma 1, M′ is of s-binary rank at
301
+ most d. We now show that r(M′) · c(M′) =
302
+ � d
303
+ ≤s
304
+
305
+ 2d.
306
+ Since the identity d × d matrix Id is a sub-matrix of L, we have that NId = N is (0, 1)-matrix
307
+ and a sub-matrix of P and therefore of M′. Therefore, r(M′) ≥ r(N) = 2d. Since Id is a sub-
308
+ matrix of N, by the same argument, c(M′) ≥ c(L) =
309
+ � d
310
+ ≤s
311
+
312
+ . Therefore r(M′) · c(M′) ≥
313
+ � d
314
+ ≤s
315
+
316
+ 2d.
317
+ Thus, r(M′) · c(M′) =
318
+ � d
319
+ ≤s
320
+
321
+ 2d.
322
+ We now show that M′ has s-binary rank d. Suppose the contrary, i.e., M′ has binary rank
323
+ d′ < d. Then there are 2d × d′ (0, 1)-matrix N and d′ ×
324
+ � d
325
+ ≤s
326
+
327
+ (0, 1)-matrix L such that P = NL
328
+ and M′[i, j] = 0 iff P[i, j] = 0. Now by Lemma 2, r(M′) ≤ r(P) ≤ r(N) ≤ 2d′ < 2d, which gives a
329
+ contradiction.
330
+ 4
331
+ Testing The s-Binary Rank
332
+ In this section, we present the adaptive and non-adaptive testing algorithms for s-binary rank at
333
+ most d. We first give the adaptive algorithm and prove Theorem 1.
334
+ 4.1
335
+ The Adaptive Tester
336
+ In this section, we prove Theorem 1.
337
+ Consider the tester Adaptive-Test-Rank in Figure 1. The tester, at every iteration of the
338
+ main While-loop (step 2) has a set X of rows of M and a set Y of columns of M. If |X| ≥ |Y |
339
+ (step 5), the tester first tries to extend M[X, Y ] with a new column (steps 6-8). If it succeeds, it
340
+ moves to the next iteration. Otherwise, it tries to extend M[X, Y ] with a new row (steps 9-12). If
341
+ it succeeds, it moves to the next iteration. Otherwise, it tries to extend M[X, Y ] with a new row
342
+ and a new column (steps 21-26). If it succeeds, it moves to the next iteration. If it fails, it accepts
343
+ (step 27). If |X| < |Y | (step 13), it starts with the row of M[X, Y ] (steps 14-16), then the column
344
+ (steps 18-20), and then both (steps 21-26). If it fails, it accepts (step 27).
345
+ If |X| · |Y | >
346
+ � d
347
+ ≤s
348
+
349
+ 2d (step 2 and then step 28) or the s-binary rank of M[X, Y ] is greater than
350
+ d (step 3), then it rejects.
351
+ We first prove
352
+ Lemma 7. Let t = 9d/ǫ. Tester Adaptive-Test-Rank makes at most 2
353
+ � d
354
+ ≤s
355
+
356
+ 2dt = ˜O
357
+ �� d
358
+ ≤s
359
+
360
+ 2d�
361
+
362
+ queries.
363
+ Proof. We prove by induction that at every iteration of the main While-loop (step 2), the tester
364
+ knows the entries of M[X, Y ], and the total number of queries, qX,Y , is at most 2|X||Y |t. Since
365
+ the While-loop condition is |X||Y | ≤
366
+ � d
367
+ ≤s
368
+
369
+ 2d, the result follows.
370
+ At the beginning of the algorithm, no queries are made, and |X| = |Y | = 1. Then 2|X||Y |t =
371
+ 2t > 0 = qX,Y .
372
+ Suppose, at the kth iteration, the tester knows the entries of M[X, Y ] and
373
+ qX,Y ≤ 2|X||Y |t. We prove the result for the (k + 1)th iteration.
374
+ We have the following cases (at the (k + 1)th iteration)
375
+ Case I. |X| ≥ |Y | (step 5) and, for some x, M[x, Y ] is a new row to M[X, Y ] (step 8).
376
+ 6
377
+
378
+ Adaptive-Test-Rank(d, s, M, n, m, ǫ)
379
+ Input: Oracle that accesses the entries of n × m (0, 1)-matrix M.
380
+ Output: Either “Accept” or “Reject”
381
+ 1. X ← {1}; Y ← {1}; t = 9d/ǫ.
382
+ 2. While |X| · |Y | ≤
383
+ � d
384
+ ≤s
385
+
386
+ 2d do
387
+ 3.
388
+ If the s-binary rank of M[X, Y ] is greater than d, then Reject.
389
+ 4.
390
+ Finish ← False; X′ ← Ø; Y ′ ← Ø. /∗ X′ and Y ′ are multi-sets.
391
+ 5.
392
+ If |X| ≥ |Y | then
393
+ 6.
394
+ While (NOT Finish) AND |X′| < t
395
+ 7.
396
+ Draw uniformly at random x ∈ [n]\X; X′ ← X′ ∪ {x};
397
+ 8.
398
+ If M[x, Y ] is a new row to M[X, Y ] then X ← X ∪ {x}; Finish ← True.
399
+ 9.
400
+ If (NOT Finish) then
401
+ 10.
402
+ While (NOT Finish) AND |Y ′| < t
403
+ 11.
404
+ Draw uniformly at random y ∈ [m]\Y ; Y ′ ← Y ′ ∪ {y}.
405
+ 12.
406
+ If M[X, y] is new column to M[X, Y ] then Y ← Y ∪ {y}; Finish ← True.
407
+ 13.
408
+ Else (|X| < |Y |)
409
+ 14.
410
+ While (NOT Finish) AND |Y ′| < t
411
+ 15.
412
+ Draw uniformly at random y ∈ [m]\Y ; Y ′ ← Y ′ ∪ {y};
413
+ 16.
414
+ If M[X, y] is a new column to M[X, Y ] then Y ← Y ∪ {y}; Finish ← True.
415
+ 17.
416
+ If (NOT Finish) then
417
+ 18.
418
+ While (NOT Finish) AND |X′| < t
419
+ 19.
420
+ Draw uniformly at random x ∈ [n]\X; X′ ← X′ ∪ {x}
421
+ 20.
422
+ If M[x, Y ] is a new row to M[X, Y ] then X ← X ∪ {x}; Finish ← True.
423
+ 21.
424
+ While (NOT Finish) AND X′ ̸= Ø do
425
+ 22.
426
+ Draw uniformly at random x ∈ X′ and y ∈ Y ′
427
+ 23.
428
+ If M[x, Y ∪ {y}] is a new row to M[X, Y ∪ {y}] OR, equivalently,
429
+ 24.
430
+ M[X ∪ {x}, y] is a new column to M[X ∪ {x}, Y ]
431
+ 25.
432
+ then X ← X ∪ {x}; Y ← Y ∪ {y}; Finish ← True.
433
+ 26.
434
+ else X′ ← X′\{x}; Y ′ ← Y ′\{y}.
435
+ 27.
436
+ If (NOT Finish) then Accept
437
+ 28.Reject
438
+ Figure 1: An adaptive tester for s-binary rank at most d.
439
+ In that case, Finish becomes true, and no other sub-while-loop is executed. Therefore, the
440
+ number of queries made at this iteration is at most |Y |t (to find all M[x, Y ]), and one element x is
441
+ added to X. Then, the tester knows all the entries of M[X ∪ {x}, Y ] and
442
+ qX∪{x},Y = qX,Y + |Y |t ≤ 2|X||Y |t + |Y |t ≤ 2|X ∪ {x}| · |Y |t,
443
+ and the result follows.
444
+ Case II. |X| ≥ |Y | (step 5), for all x′ ∈ X′, M[x′, Y ] is not a new row to M[X, Y ] (step 8), and
445
+ for some y, M[X, y] is a new column to M[X, Y ] (step 12).
446
+ 7
447
+
448
+ In that case, Finish becomes true, and no other sub-while-loop is executed after the second
449
+ sub-while-loop (step 10).
450
+ Therefore, in this case, the number of queries made at this iteration is at most |Y |t + |X|t.
451
+ |X|t queries in the first sub-while-loop (to find M[x, Y ] for all x ∈ X′), and at most |Y |t queries
452
+ in the second sub-while-loop (to find M[X, y′] for all y′ ∈ Y ′). Then one element y is added to Y .
453
+ Therefore, the tester knows the entries of M[X, Y ∪ {y}] and, since |Y | ≤ |X|,
454
+ qX,Y ∪{y} = qX,Y + |X|t + |Y |t ≤ 2|X||Y |t + 2|X|t = 2|X| · |Y ∪ {y}|t,
455
+ and the result follows.
456
+ Case III. |X| ≥ |Y |, for all x′ ∈ X′, M[x′, Y ] is not a new row to M[X, Y ], for all y′ ∈ Y ′, M[X, y′]
457
+ is not a new column to M[X, Y ], and for some x ∈ X′, y ∈ Y ′, M[x, Y ∪ {y}] is a new row to
458
+ M[X, Y ∪ {y}] (step 23).
459
+ In this case, |X′| = |Y ′| = t, the number of queries is |X|t + |Y |t + t. Exactly |X|t queries in
460
+ the first sub-while-loop, |Y |t queries in the second sub-while-loop, and at most5 t queries in the
461
+ sub-while-loop in step 21. Then one element x is added to X, and one element y is added to Y .
462
+ Then the tester knows the entries of M[X ∪ {x}, Y ∪ {y}] and
463
+ qX∪{x},Y ∪{y} = qX,Y + |X|t + |Y |t + t ≤ 2|X| · |Y |t + |X|t + |Y |t + t ≤ 2|X ∪ {x}| · |Y ∪ {y}|t.
464
+ Case IV. |X| ≥ |Y |, for all x′ ∈ X′, M[x′, Y ] is not a new row to M[X, Y ], for all y′ ∈ Y ′, M[X, y′]
465
+ is not a new column to M[X, Y ], and for all the drawn pairs x ∈ X′, y ∈ Y ′, M[x, Y ∪ {y}] is not
466
+ a new row to M[X, Y ∪ {y}] (step 23).
467
+ In this case, Finish will have value False, and the tester accepts in step 27.
468
+ The analysis of the case when |X| < |Y | is similar to the above analysis.
469
+ We now prove the completeness of the tester.
470
+ Lemma 8. If M is a n × m (0, 1)-matrix of s-binary rank at most d, then the tester Adaptive-
471
+ Test-Rank accepts with probability 1.
472
+ Proof. The tester rejects if and only if one of the following occurs,
473
+ 1. M[X, Y ] has s-binary rank greater than d.
474
+ 2. |X| · |Y | >
475
+ � d
476
+ ≤s
477
+
478
+ 2d.
479
+ If M[X, Y ] has s-binary rank greater than d, then M has s-binary rank greater than d. This is
480
+ because, if M = NL, then M[X, Y ] = N[X, [d]] · L[[d], Y ]. So item 1 cannot occur.
481
+ Before we show that item 2 cannot occur, we prove the following:
482
+ Claim 1. The rows (resp. columns) of M[X, Y ] are distinct.
483
+ Proof. The steps in the tester where we add rows or columns are steps 8, 12 16, 20, and 23. In
484
+ steps 8, 12 16, 20 it is clear that a row (resp. column) is added only if it is a new row (resp.
485
+ column) to M[X, Y ]. Consider step 23 and suppose, w.l.o.g |X| ≥ |Y |. This step is executed only
486
+ when Finish = False. This happens when |X′| = |Y ′| = t, for every x ∈ X′, M[x, Y ] is not a
487
+ new row to M[X, Y ], and for every y ∈ Y ′, M[X, y] is not a new column to M[X, Y ]. Then x
488
+ 5This is because, for x ∈ X′, y ∈ Y ′, the tester already knows M[x, Y ] and M[X, y] from the first and second
489
+ sub-while-loop and only needs to query M[x, y].
490
+ 8
491
+
492
+ and y are added to X and Y , respectively, if M[x, Y ∪ {y}] is a new row to M[X, Y ∪ {y}]. Then,
493
+ by Lemma 3, M[X ∪ {x}, y] is a new column to M[X ∪ {x}, Y ]. So, the rows (and columns) in
494
+ M[X ∪ {x}, Y ∪ {y}] are distinct. This implies the result.
495
+ Suppose, to the contrary, |X| · |Y | >
496
+ � d
497
+ ≤s
498
+
499
+ 2d. Since M′ = M[X, Y ] satisfies r(M′)c(M′) =
500
+ |X| · |Y | >
501
+ � d
502
+ ≤s
503
+
504
+ 2d, by Lemma 4, the s-binary rank of M′, and therefore of M, is greater than d. A
505
+ contradiction.
506
+ We now prove the soundness of the tester.
507
+ We first prove the following.
508
+ Claim 2. Let M be a n×m (0, 1)-matrix, X ⊆ [n], and Y ⊆ [m]. Suppose there are two functions,
509
+ ′ : [n] → X and ′′ : [m] → Y , such that
510
+ 1. For every x ∈ [n], M[x, Y ] = M[x′, Y ].
511
+ 2. For every y ∈ [m], M[X, y] = M[X, y′′].
512
+ 3. For every x ∈ [n] and y ∈ [m], M[x, y] = M[x′, y′′].
513
+ Then M has at most |X| distinct rows and |Y | distinct columns, and its s-binary rank is the s-binary
514
+ rank of M[X, Y ].
515
+ Proof. Let x ∈ [n]\X. For every y, M[x, y] = M[x′, y′′] = M[x′, y]. Therefore, row x in M is equal
516
+ to row x′. Similarly, column y in M is equal to column y′′.
517
+ Since adding equal columns and rows to a matrix does not change the s-binary rank6, we have
518
+ brs(M[X, Y ]) = brs(M[X, [m]]) = brs(M).
519
+ The following Claim is proved in [7] (Claim 10). Here, we give the proof for completeness.
520
+ Claim 3. Let M be a (0, 1)-matrix that is ǫ-far from having s-binary rank at most d. Let X ⊆ [n]
521
+ and Y ⊆ [m], such that brs(M[X, Y ]) ≤ d, the columns of M[X, Y ] are distinct, and the rows of
522
+ M[X, Y ] are distinct. Then one of the following must hold:
523
+ 1. The number of rows x ∈ [n] where M[x, Y ] is a new row to M[X, Y ] is at least nǫ/3.
524
+ 2. The number of columns y ∈ [m] where M[X, y] is a new column to M[X, Y ] is at least mǫ/3.
525
+ 3. The number pairs (x, y), x ̸∈ X, y ̸∈ Y , where, M[x, Y ] = M[x′, Y ] for some x′ ∈ X,
526
+ M[X, y] = M[X, y′′] for some y′′ ∈ Y , and M[x, y] ̸= M[x′, y′′], is at least mnǫ/3.
527
+ Proof. Assume, to the contrary, that none of the above statements holds. Change every row x in
528
+ M where M[x, Y ] is a new row to M[X, Y ] to a zero row. Let X′ be the set of such rows. Change
529
+ every column y in M where M[X, y] is a new row to M[X, Y ] to a zero column. Let Y ′ be the
530
+ set of such columns. For every other entry (x, y), x ̸∈ X, y ̸∈ Y that is not changed to zero and
531
+ M[x, y] ̸= M[x′, y′′], change M[x, y] to M[x′, y′′]. Let M′ be the matrix obtained from the above
532
+ changes.
533
+ The number of entries (x, y) where M[x, y] ̸= M′[x, y] is less than (nǫ/3)m + (mǫ/3)n +
534
+ mnǫ/3 = ǫmn. Therefore, M′ is ǫ-close to M. By claim 3, brs(M′) = brs(M[[n]\X′, [m]\Y ′]) =
535
+ brs(M[X, Y ]) ≤ d. A contradiction.
536
+ 6If we add a column to a matrix that is equal to column y, then the rectangles that cover column y can be extended
537
+ to cover the added column.
538
+ 9
539
+
540
+ We now prove the completeness of the tester.
541
+ Lemma 9. If M is ǫ-far from having s-binary rank d, then with probability at least 2/3, Adaptive-
542
+ Test-Rank rejects.
543
+ Proof. Consider the while-loop in step 2 at some iteration i. If brs(M[X, Y ]) > d, then the tester
544
+ rejects in step 3. We will now show that if brs(M[X, Y ]) ≤ d, then, with probability at most 3e−2d,
545
+ the tester accepts at iteration i.
546
+ To this end, let brs(M[X, Y ]) ≤ d. Then, by Claim 3, one of the following holds.
547
+ 1. The number of rows x ∈ [n] where M[x, Y ] is a new row to M[X, Y ] is at least nǫ/3.
548
+ 2. The number of columns y ∈ [m] where M[X, y] is a new column to M[X, Y ] is at least mǫ/3.
549
+ 3. The number pairs (x, y), x ̸∈ X, y ̸∈ Y , where, M[x, Y ] = M[x′, Y ] for some x′ ∈ X,
550
+ M[X, y] = M[X, y′′] for some y′′ ∈ Y , and M[x, y] ̸= M[x′, y′′], is at least mnǫ/3.
551
+ Now at the ith iteration, suppose w.l.o.g, |X| ≥ |Y | (the other case |Y | < |X| is similar). If item 1
552
+ occurs, then with probability at least p = 1 − (1 − ǫ/3)t ≥ 1 − e−2d, the tester finds a new row
553
+ to M[X, Y ] and does not accept at iteration i. If item 2 occurs, then if it does not find a new
554
+ row to M[X, Y ], with probability at least p, the tester finds a new column to M[X, Y ] and does
555
+ not accept. If item 3 occurs, and it does not find a new row or column to M[X, Y ], then with
556
+ probability at least p, it finds such a pair and does not accept. Therefore, with probability at most
557
+ 3(1 − p) ≤ 3e−2d, the tester accepts at iteration i.
558
+ Since the while-loop runs at most |X| + |Y | ≤ 2|X||Y | ≤ 2
559
+ � d
560
+ ≤s
561
+
562
+ 2d ≤ 22d+1 iterations, with
563
+ probability at most 3e−2d22d+1 ≤ 1/3, the tester accepts in while-loop. Therefore, with proba-
564
+ bility at least 2/3, the tester does not accept in the while-loop. Thus, it either rejects because
565
+ brs(M[X, Y ]) > d or rejects in step 28.
566
+ 4.2
567
+ The Non-Adaptive Tester
568
+ In this section, we prove Theorem 2.
569
+ First, consider Adaptive-Test-Rank in Figure 1. Consider steps 7,11,15, and 19, where it
570
+ draws a new column or row. We prove.
571
+ Lemma 10. Let t = 9d/ǫ. At each iteration of Adaptive-Test-Rank, the total number of uni-
572
+ formly at random rows x ∈ [n] drawn is at most (|X| + min(|X|, |Y | − 1))t, and the number of
573
+ uniformly at random rows y ∈ [m] drawn is at most (|Y | + min(|X|, |Y |))t.
574
+ Proof. We prove by induction that at every iteration of the main While-loop (step 2), the total
575
+ number of random rows drawn by the tester, nX,Y , is at most (|X| + min(|X|, |Y | − 1))t, and the
576
+ total number of random columns drawn, mX,Y , is at most (|Y | + min(|X|, |Y |))t.
577
+ At the beginning, |X| = |Y | = 1, and the number of columns and rows is 1. In that case,7,
578
+ nX,Y = 1 ≤ t and mX,Y = 1 ≤ 2t. Suppose, at the kth iteration, the induction statement is true.
579
+ We prove the result for the (k + 1)th iteration.
580
+ At the (k + 1)th iteration, we have the following cases.
581
+ Case I. |X| ≥ |Y | (step 5) and, for some x, M[x, Y ] is a new row to M[X, Y ] (step 8).
582
+ 7We assume that the first column/row drawn is column/row one
583
+ 10
584
+
585
+ Non-Adaptive-Test-Rank(d, s, M, n, m, ǫ)
586
+ Input: Oracle that accesses the entries of (0, 1)-matrix M.
587
+ Output: Either “Accept” or “Reject”.
588
+ 1.
589
+ T ←
590
+ 324·d2( d
591
+ ≤s)2d
592
+ ǫ2
593
+ .
594
+ 2.
595
+ Dray uniformly at random x(1), . . . , x(T) ∈ [n].
596
+ 3.
597
+ Dray uniformly at random y(1), . . . , y(T) ∈ [m].
598
+ 4.
599
+ For every i ∈ [T] and j ∈ [T] such that i · j ≤ T
600
+ 5.
601
+ D[i, j] ← Query M[x(i), y(j)]
602
+ 6.
603
+ u = 1; w = 1.
604
+ 7.
605
+ Run Adaptive-Test-Rank(d, s, M, n, m, ǫ)
606
+ When the tester asks for a uniform at random x - return x(u); u ← u + 1
607
+ When the tester asks for a uniform at random y - return y(w); w ← w + 1
608
+ When the tester makes the Query M[x(i), y(j)] - return D[i, j]
609
+ Figure 2: A non-adaptive tester for s-binary rank at most d.
610
+ In that case, Finish becomes true, and no other sub-while-loop is executed. Therefore, the
611
+ number of rows drawn at this iteration is at most t, and one element x is added to X. No columns
612
+ are drawn. Then,
613
+ nX∪{x},Y ≤ nX,Y + t ≤ (|X| + min(|X|, |Y | − 1) + 1)t ≤ (|X ��� {x}| + min(|X ∪ {x}|, |Y | − 1))t,
614
+ and
615
+ mX∪{x},Y = mX,Y ≤ (|Y | + min(|X|, |Y |))t ≤ (|Y | + min(|X ∪ {x}|, |Y |))t.
616
+ Thus, the result follows for this case.
617
+ Case II. |X| ≥ |Y | (step 5), for all x′ ∈ X′, M[x′, Y ] is not a new row to M[X, Y ] (step 8), and
618
+ for some y, M[X, y] is a new column to M[X, Y ] (step 12).
619
+ In that case, Finish becomes true, and no other sub-while-loop is executed after the second
620
+ sub-while-loop (step 10).
621
+ Therefore, in this case, the number of rows drawn at this iteration is t, one element y is added
622
+ to Y , and the number of columns drawn is at most t. Then
623
+ nX,Y ∪{y} = nX,Y + t
624
+
625
+ (|X| + min(|X|, |Y | − 1) + 1)t
626
+ =
627
+ (|X| + |Y |)t = (|X| + min(|X|, |Y ∪ {y}| − 1))t,
628
+ and
629
+ mX,Y ∪{y} ≤ mX,Y + t ≤ (|Y | + min(|X|, |Y |) + 1)t ≤ (|Y ∪ {y}| + min(|X|, |Y ∪ {y}|))t.
630
+ Thus, the result follows for this case.
631
+ Case III. |X| < |Y | (step 13), and for some y, M[X, y] is a new column to M[X, Y ] (step 16).
632
+ 11
633
+
634
+ In that case, Finish becomes true, and no other sub-while-loop is executed. Therefore, the
635
+ number of columns drawn at this iteration is at most t, and one element y is added to Y . No rows
636
+ are drawn. Then,
637
+ nX,Y ∪{y} = nX,Y ≤ (|X| + min(|X|, |Y | − 1))t ≤ (|X| + min(|X|, |Y ∪ {y}| − 1))t,
638
+ and
639
+ mX,Y ∪{y} ≤ mX,Y + t ≤ (|Y | + min(|X|, |Y |) + 1)t = (|Y ∪ {y}| + min(|X|, |Y ∪ {y}|))t.
640
+ Thus, the result follows for this case.
641
+ Case IV. |X| < |Y | (step 13), for all y′ ∈ Y ′, M[X, y′] is not a new row to M[X, Y ], and for some
642
+ x, M[x, Y ] is a new column to M[X, Y ] (step 20). In that case, Finish becomes true, and no other
643
+ sub-while-loop is executed after the fourth sub-while-loop (step 18).
644
+ In this case, the number of rows drawn at this iteration is t, one element x is added to X, and
645
+ the number of columns drawn is at most t. Then
646
+ nX∪{x},Y = nX,Y + t
647
+
648
+ (|X| + min(|X|, |Y | − 1) + 1)t
649
+
650
+ (|X ∪ {x}| + min(|X ∪ {x}|, |Y | − 1))t
651
+ mX∪{x},Y ≤ mX,Y + t ≤ (|Y | + min(|X|, |Y |) + 1)t = (|Y | + min(|X ∪ {x}|, |Y |))t.
652
+ Thus, the result follows for this case.
653
+ Case V. For all x′ ∈ X′, M[x′, Y ] is not a new row to M[X, Y ], for all y′ ∈ Y ′, M[X, y′] is not a
654
+ new column to M[X, Y ], and for some x ∈ X′, y ∈ Y ′, M[x, Y ∪{y}] is a new row to M[X, Y ∪{y}]
655
+ (step 23).
656
+ In this case, the number of rows drawn at this iteration is t, the number of columns drawn is t,
657
+ one element x is added to X, and one element y is added to Y . Then
658
+ nX∪{x},Y ∪{y} = nX,Y + t
659
+
660
+ (|X| + min(|X|, |Y | − 1) + 1)t
661
+
662
+ (|X ∪ {x}| + min(|X ∪ {x}|, |Y ∪ {y}| − 1))t.
663
+ mX∪{x},Y ∪{y} = mX,Y + t
664
+
665
+ (|Y | + min(|X|, |Y |) + 1)t
666
+
667
+ (|Y ∪ {y}| + min(|X ∪ {x}|, |Y ∪ {y}|))t.
668
+ We are now ready to prove Theorem 2.
669
+ Proof. By Lemma 10, the total number of rows and columns drawn in Adaptive-Test-Rank
670
+ up to iteration t is at most n′ := 9(|X| + min(|X|, |Y | − 1))d/ǫ ≤ 18|X|d/ǫ and m′ := 9(|Y | +
671
+ min(|X|, |Y |)d/ǫ ≤ 18|Y |d/ǫ, respectively. We also have |X| · |Y | ≤
672
+ � d
673
+ ≤s
674
+
675
+ 2d. So
676
+ n′ · m′ ≤ 324|X||Y |d2/ǫ2 ≤ T :=
677
+ 324 · d2� d
678
+ ≤s
679
+
680
+ 2d
681
+ ǫ2
682
+ .
683
+ Consider the tester Non-Adaptive-Test-Rank in Figure 2. The tester draws T rows x(1), . . . ,
684
+ x(T) ∈ [n], and columns y(1), . . . , y(T) ∈ [m] and queries all M[x(i), y(j)] where ij ≤ T and puts the
685
+ 12
686
+
687
+ result in the table D. Then it runs Adaptive-Test-Random using the above-drawn rows and
688
+ columns. We now show that all the queries that Adaptive-Test-Random makes can be fetched
689
+ from the table D.
690
+ At any iteration, the number of rows drawn is at most n′, and the number of rows drawn is at
691
+ most m′. Therefore, the tester needs to know (in the worst case) all the entries M[x(i), y(j)] where
692
+ i ≤ n′ and j ≤ m′. Since ij ≤ n′m′ ≤ T, the result follows.
693
+ The number of queries that the tester makes is
694
+ T
695
+
696
+ i=1
697
+ T
698
+ i = O(T ln T) = ˜O
699
+ �� d
700
+ ≤s
701
+
702
+ 2d
703
+ ǫ2
704
+
705
+ .
706
+ 5
707
+ Testing the Exact s-Binary Rank
708
+ We first prove the following.
709
+ Lemma 11. Let M and M′ be n × m (0, 1)-matrices that differ in one row (or column). Then
710
+ |brs(M) − brs(M′)| ≤ 1.
711
+ Proof. Suppose brs(M) = d and M′ differ from M in row k. Let N and L be n × d (0, 1)-matrix
712
+ and d × m (0, 1)-matrix, respectively, such that P = NL, for every (i, j) ∈ [n] × [m], P[i, j] ≤ s,
713
+ and P[i, j] = 0 if and only if M[i, j] = 0. Add to N a column (as a (d + 1)th column) that all its
714
+ entries are zero except the k-th entry, which equals 1. Then change N[k, j] to zero for all j ∈ [d].
715
+ Let N ′ be the resulting matrix. Add to L another row (as a (d + 1)th row) equal to the k-th row
716
+ of M′. Let L′ be the resulting matrix. Let P ′ = N ′L′. It is easy to see that P ′[i, j] = P[i, j] for all
717
+ i ̸= k and j, and the kth row of P ′ is equal to the kth row of M′. Then, for every (i, j) ∈ [n] × [m],
718
+ P ′[i, j] ≤ s, and P ′[i, j] = 0 if and only if M′[i, j] = 0. Therefore, brs(M′) ≤ d + 1 = brs(M) + 1.
719
+ In the same way, brs(M) ≤ brs(M′) + 1.
720
+ Lemma 12. Let η = d2/(nm). Let M be n × m (0, 1)-matrix. If M is ǫ-close to having s-binary
721
+ rank at most d, then M is (ǫ + η)-close to having s-binary rank d.
722
+ Proof. We will show that for every n × m (0, 1)-matrix H of s-binary rank at most d − 1, there is a
723
+ n × m (0, 1)-matrix G of s-binary rank d that is η-close to H. Therefore, if M is ǫ-close to having
724
+ s-binary rank at most d, then it is (ǫ + η)-close to having s-binary rank d.
725
+ Define the n×m (0, 1)-matrices Gk, k ∈ [d]∪{0}, where G0 = H and for k ≥ 1, Gk[i, j] = H[i, j]
726
+ if j > k or i > d, and Gk[[d], [k]] = Id[[d], [k]] where Id is the d × d identity matrix.
727
+ Since
728
+ Gd[[d], [d]] = Id, we have brs(Gd) ≥ d. It is clear that for every k ∈ [d] ∪ {0}, Gk is (d2/nm)-close
729
+ to H. If brs(Gd) = d, then take G = Gd, and we are done. Otherwise, suppose brs(Gd) > d.
730
+ Now consider a sequence H = G0, G1, G2, . . . , Gd. By Lemma 11, we have brs(Gi−1) − 1 ≤
731
+ brs(Gi) ≤ brs(Gi−1) + 1. Now since brs(G0) = brs(H) ≤ d − 1 and brs(Gd) > d, by the discrete
732
+ intermediate value theorem, there must be k ∈ [d] such that brs(Gk) = d. Then take G = Gk, and
733
+ we are done.
734
+ Now, the tester for testing the s-binary rank d runs as follows.
735
+ If mn < 2d2/ǫ, then find
736
+ all the entries of M with mn < 2d2/ǫ queries. If brs(M) = d, then accept. Otherwise, reject.
737
+ 13
738
+
739
+ If mn ≥ 2d2/ǫ, then run Adaptive-Test-Rank(d, s, M, n, m, ǫ/2) (for the non-adaptive, we run
740
+ Non-Adaptive-Test-Rank(d, s, M, n, m, ǫ/2)) and output its answer.
741
+ We now show the correctness of this algorithm. If M is of s-binary rank d, then it is of s-binary
742
+ rank at most d, and the tester accepts.
743
+ Now, suppose f is ǫ-far from having s-binary rank d. If mn < 2d2/ǫ, the tester rejects. If
744
+ mn ≥ 2d2/ǫ, then, by Lemma 12, f is (ǫ − η)-far from having s-binary rank at most d, where
745
+ η = d2/(nm). Since η = d2/(nm) ≤ ǫ/2, the function f is (ǫ/2)-far from having s-binary rank at
746
+ most d, and therefore the tester, with probability at least 2/3, rejects.
747
+ References
748
+ [1] Maria-Florina Balcan, Yi Li, David P. Woodruff, and Hongyang Zhang. Testing matrix rank,
749
+ optimally. In Proceedings of the Thirtieth Annual ACM-SIAM Symposium on Discrete Algo-
750
+ rithms, SODA 2019, San Diego, California, USA, January 6-9, 2019, pages 727–746, 2019.
751
+ doi:10.1137/1.9781611975482.46.
752
+ [2] Parinya Chalermsook, Sandy Heydrich, Eugenia Holm, and Andreas Karrenbauer. Nearly tight
753
+ approximability results for minimum biclique cover and partition. In Andreas S. Schulz and
754
+ Dorothea Wagner, editors, Algorithms - ESA 2014 - 22th Annual European Symposium, Wro-
755
+ claw, Poland, September 8-10, 2014. Proceedings, volume 8737 of Lecture Notes in Computer
756
+ Science, pages 235–246. Springer, 2014. doi:10.1007/978-3-662-44777-2\_20.
757
+ [3] Dana Ron. Private Communication.
758
+ [4] David A. Gregory, Norman J. Pullman, Kathryn F. Jones, and J. Richard Lundgren. Biclique
759
+ coverings of regular bigraphs and minimum semiring ranks of regular matrices. J. Comb. Theory,
760
+ Ser. B, 51(1):73–89, 1991. doi:10.1016/0095-8956(91)90006-6.
761
+ [5] Yi Li, Zhengyu Wang, and David P. Woodruff.
762
+ Improved testing of low rank matrices.
763
+ In The 20th ACM SIGKDD International Conference on Knowledge Discovery and Data
764
+ Mining, KDD ’14, New York, NY, USA - August 24 - 27, 2014, pages 691–700, 2014.
765
+ doi:10.1145/2623330.2623736.
766
+ [6] Yonatan Nakar and Dana Ron.
767
+ On the testability of graph partition properties.
768
+ In
769
+ Eric Blais, Klaus Jansen, Jos´e D. P. Rolim, and David Steurer, editors, Approxima-
770
+ tion, Randomization, and Combinatorial Optimization. Algorithms and Techniques, AP-
771
+ PROX/RANDOM 2018,
772
+ August 20-22,
773
+ 2018
774
+ - Princeton,
775
+ NJ, USA, volume
776
+ 116
777
+ of
778
+ LIPIcs,
779
+ pages
780
+ 53:1–53:13.
781
+ Schloss
782
+ Dagstuhl
783
+ -
784
+ Leibniz-Zentrum
785
+ f¨ur
786
+ Informatik,
787
+ 2018.
788
+ doi:10.4230/LIPIcs.APPROX-RANDOM.2018.53.
789
+ [7] Michal Parnas, Dana Ron, and Adi Shraibman. Property testing of the boolean and binary
790
+ rank. Theory Comput. Syst., 65(8):1193–1210, 2021. doi:10.1007/s00224-021-10047-8.
791
+ [8] Jir´ı Sgall. Bounds on pairs of families with restricted intersections. Comb., 19(4):555–566, 1999.
792
+ doi:10.1007/s004939970007.
793
+ 14
794
+
3dE3T4oBgHgl3EQfPwmd/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
49AyT4oBgHgl3EQfcPf_/content/tmp_files/2301.00281v1.pdf.txt ADDED
@@ -0,0 +1,305 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ arXiv:2301.00281v1 [cs.LG] 31 Dec 2022
2
+ arXiv® 2022 (cs.LG) 1-7
3
+ Submitted 12/22; Published 12/22
4
+ Lightmorphic Signatures Analysis Toolkit
5
+ Dumitru Damian
6
7
+ Information and Communication Engineering
8
+ Research and development consultant
9
+ Timis,oara, RO
10
+ Abstract
11
+ In this paper we discuss the theory used in the design of an open source lightmorphic sig-
12
+ natures analysis toolkit (LSAT). In addition to providing a core functionality, the software
13
+ package enables specific optimizations with its modular and customizable design.
14
+ To promote its usage and inspire future contributions, LSAT is publicly available. By
15
+ using a self-supervised neural network and augmented machine learning algorithms, LSAT
16
+ provides an easy-to-use interface with ample documentation.
17
+ The experiments demonstrate that LSAT improves the otherwise tedious and error-
18
+ prone tasks of translating lightmorphic associated data into usable spectrograms, enhanced
19
+ with parameter tuning and performance analysis.
20
+ With the provided mathematical functions, LSAT validates the nonlinearity encoun-
21
+ tered in the data conversion process while ensuring suitability of the forecasting algorithms.
22
+ Keywords:
23
+ lightmorphic, machine learning, spectrogram, graph chord, neural network
24
+ 1. Introduction
25
+ It is common knowledge, in the machine learning domain, to use differential values, since
26
+ they provide a simple way to model the data. However, such algorithms may not fit the
27
+ lightmorphic signature properly, leading to a reduced quality of the obtained results. Train-
28
+ ing a neural network to predict the lightmorphic signature can significantly increase the data
29
+ quality. This is the task that LSAT tries to accomplish.
30
+ As such we define the lightmorphic metric learning (LML) as a branch of machine
31
+ learning algorithms, set out with the purpose of learning lightmorphic signatures from
32
+ multiple datasets trough usage of vibrating graph chords.
33
+ In the pursuing sections we describe the main features of the toolkit, explain the general
34
+ mathematical concepts and finally detail the plans regarding future functionalities.
35
+ 2. General mathematical concepts
36
+ In this section we expand the mathematical concepts and link them with the reasoning
37
+ encountered in the implemented code.
38
+ We define the lightmorphic signature as a function of: light intensity (I) that varies
39
+ according to seasons and local weather conditions, trajectory distribution characteristics
40
+ ©2022 Dumitru Damian.
41
+ License: CC-BY 4.0, see https://creativecommons.org/licenses/by/4.0/.
42
+ Typeset in LaTex using the JMLR LaTeX style file https://jmlr.org/author-info.html.
43
+
44
+ Damian
45
+ (D), and specific adjustments (T):
46
+ fL⊙ =
47
+ I
48
+
49
+ 1
50
+ D
51
+
52
+ 1
53
+ T
54
+
55
+ 1
56
+ Γtζtdt
57
+ (1)
58
+ where:
59
+ • Γt – trajectory tensor
60
+ • ζt – point in time specificity
61
+ Storage of these trajectory specific lightmorphic signatures is done in a database (Θ).
62
+ The segments containing isochronous surfaces with similarities are stored in another database
63
+ (Φ) that serves as a baseline for training the neural network implementation.
64
+ The isochronous surfaces that constitute the lightmorphic signature are interlinked
65
+ trough the definition and usage of graph chords (δ(t)). Observing their vibrational am-
66
+ plitude allows the prediction of alternative lightmorphic signatures and, at the same time,
67
+ correction of the already known values.
68
+ Since the primary light source considered is the Earth’s Sun, specific spacetime metrics
69
+ (ex. gµν, ηµν, h+, h×, Gµν) have to be used in order to describe the encountered anisotropies.
70
+ These are implemented as a function of distant astrophysical forces that stretch and com-
71
+ press the fabric of spacetime.
72
+ According to special relativity, spacetime is seen as a four dimensional manifold de-
73
+ scribed by a flat Minkowski metric defined in Cartesian coordinates (t, x, y, z, c = 1)
74
+ as:
75
+ ηµν =
76
+
77
+
78
+
79
+
80
+ −1
81
+ 0
82
+ 0
83
+ 0
84
+ 0
85
+ 1
86
+ 0
87
+ 0
88
+ 0
89
+ 0
90
+ 1
91
+ 0
92
+ 0
93
+ 0
94
+ 0
95
+ 1
96
+
97
+
98
+
99
+  ,
100
+ (2)
101
+ When considering the geometry of curved space, we have made use of the metric gµν,
102
+ that replaces the flat Minkowski metric ηµν. This substitution was done considering that
103
+ the geometry of curved space will eventually reduce to the flat spacetime of special relativity
104
+ at a sufficiently small scale.
105
+ The interaction between curvature of spacetime and the mass distribution was modeled
106
+ following (Blackburn (2010)) work, as:
107
+ Gµν = kTµν
108
+ (3)
109
+ where Gµν is defined as the Einstein curvature tensor, Tµν is the stress-energy tensor and
110
+ represents the mass-energy distribution, while k describes the Einstein constant of gravita-
111
+ tion defined as:
112
+ k = 8πG
113
+ c4
114
+ (4)
115
+ where c is the speed of light in a vacuum.
116
+ 2
117
+
118
+ Lightmorphic Signatures Analysis Toolkit
119
+ At the same time, in order to improve the results quality, the Einstein tensor was also
120
+ considered under the form:
121
+ Gµν = Rµν − 1
122
+ 2gµνR,
123
+ (5)
124
+ where Rµν is the Riemann tensor for the local spacetime, and R is the Ricci scalar.
125
+ Since there is not one general solution for the complex Einstein equations, but a large
126
+ variety of possible solutions that apply to particular circumstances, we’ve considered a weak-
127
+ field approximation, where the nonlinear Einstein equations where approximated towards
128
+ linearity.
129
+ For example, a very small perturbation specific to a gravitational wave, will impact the
130
+ flat spacetime and it is defined as hµν(x) and it’s value will be |hµν| << 1.
131
+ Thus, the Einstein equation becomes:
132
+ gµν(x) = ηµν + hµν(x).
133
+ (6)
134
+ or by simply considering the induced strain variations:
135
+ □hµν(x) = 0,
136
+ (7)
137
+ By further pursuing such linearization, we can represent in the TT gauge, a propagating
138
+ wave, under the following form:
139
+ hTT
140
+ µν =
141
+
142
+
143
+
144
+
145
+ 0
146
+ 0
147
+ 0
148
+ 0
149
+ 0
150
+ h+
151
+
152
+ 0
153
+ 0
154
+
155
+ −h+
156
+ 0
157
+ 0
158
+ 0
159
+ 0
160
+ 0
161
+
162
+
163
+
164
+  ,
165
+ (8)
166
+ where the constant amplitudes (h+, h×) represent the two gravitational wave polariza-
167
+ tions, the plus- and cross-polarization.
168
+ We represent the distance between two neighboring points as defined by (Berit (2013))
169
+ for a flat spacetime, trough the following expression:
170
+ ds2 = −c2dt2 + dx2 + dy2 + dz2 = −c2dt2 + [1 + h+(t)]dx2 + [1 − h+(t)]dy2
171
+ That allows us to model in the TT gauge, the gravitational wave stretching along the x
172
+ axis and compression along the y axis with the specific factor of:
173
+
174
+ 1 ± h+(t) ≃ 1 + 1
175
+ 2h+(t)
176
+ Having modeled the photon’s traveling path in outer space, in order to simplify the
177
+ inherent path inhomogeneities, we separated the domains into outer space domain, atmo-
178
+ spheric domain and Earth specific domains (lithosphere, hydrosphere, biosphere, noises,
179
+ etc).
180
+ We further define the phase of an electromagnetic wave of frequency ω0 as φ. Following
181
+ Driggers (2015)’s work, we consider that the starting light phase is at 0 and it travels at
182
+ the speed of light c. After a distance L it will have a phase δφspace that can be expressed as
183
+ a distance integral over the spacetime metric,
184
+ δφspace = ω0
185
+ c
186
+ � L
187
+ 0
188
+ gdx,
189
+ (9)
190
+ 3
191
+
192
+ Damian
193
+ with g(t) = η+h(t), where η is the Minkowski metric and h(t) is the dimensionless spacetime
194
+ strain.
195
+ Summing the light phase shift δφatm and the δφEarth which is derived from the noise
196
+ sources like seismic or electromagnetic interferences, leads to the dataset of trajectory spe-
197
+ cific lightmorphic signatures:
198
+ ΦΓIDT =
199
+ N
200
+
201
+ j=1
202
+ Γj
203
+ IDT
204
+ (10)
205
+ The signature parameter estimation is performed considering a prior distribution p(Φ|L⊙)
206
+ that is updated upon receiving the new data d to give a posterior distribution p(Φ|d, L⊙)
207
+ p(Φ|d, L⊙) = p(Φ|L⊙)p(d|Φ, L⊙)
208
+ p(d|L⊙)
209
+ (11)
210
+ While observing the distribution of multiple light segments within the dataset ΦΓIDT ,
211
+ it will be possible to estimate the probability for trajectory specific lightmorphic evolution:
212
+ pΦ = f(ρk · pΦk)
213
+ (12)
214
+ where pΦk is the database’s k-th segment specific probability, ρk is the prediction weight
215
+ for the k-th segment.
216
+ 3. Software package design
217
+ The distribution matrices specific to the isochronous segmentation surfaces, which define
218
+ the lightmorphic signature model, form the LSAT core.
219
+ As such we’ve used a design principle that ensures simplicity for the whole package,
220
+ while making the source codes easy to read and maintain. As the toolkit is written in a
221
+ modular way, new functionalities can be easily plugged in. This makes the LSAT not only
222
+ a lightmorphic signature machine learning tool but also an experimental platform.
223
+ LSAT comes with plenty of documentation for all the interface functionalities and related
224
+ data structures. The README file describes the installation process and interface usage.
225
+ For developers who use the toolkit in their applications, the API documentation can provide
226
+ additional information related to functionality calls.
227
+ 4. Practical Usage
228
+ In the examples, we provide sample values for the lightmorphic signature updates, as a
229
+ function of δφatm derived by the neural network from the values of a large dataset of at-
230
+ mospheric meteorological data for 317 cities in Romania, with hundreds of thousands data
231
+ points.
232
+ Automatic learning is supported trough API calls to the domain specific data
233
+ providers.
234
+ Beyond this simple way of running the lightmorphic signatures analysis toolkit, there are
235
+ several enhancement options for advanced usage. As example, one may activate additional
236
+ functionalities that consider input parameters like complex space weather forecasting, dif-
237
+ ferent electromagnetic wave disturbances or lithosphere, hydrosphere and biosphere specific
238
+ localized data.
239
+ 4
240
+
241
+ Lightmorphic Signatures Analysis Toolkit
242
+ 5. Conclusion and Future Work
243
+ With the lightmorphic signatures analysis toolkit we provided an open source SW package
244
+ that is simple and easy-to-use.
245
+ Experiments and analysis conclude that the modular design and customization support
246
+ are performing excellent in practice and can provide the base for additional research on
247
+ lightmorphic signatures.
248
+ The toolkit is constantly being improved by new research results and user feed-back
249
+ with the ultimate goal of having an automated toolkit to use in maintaining and updating
250
+ a large database of high-quality light signatures.
251
+ Future work will focus on probability estimates, additional functionalities that mitigate
252
+ the large uncertainties in the available observational input data which arise from the complex
253
+ interaction processes. In addition, the inclusion of artificial intelligence (AI) options will
254
+ be considered while building a national/international network for lightmorphic signature
255
+ analysis.
256
+ 5
257
+
258
+ Damian
259
+ 6. References
260
+ References
261
+ Rana Adhikari. Sensitivity and noise analysis of 4 km laser interferometric gravitational
262
+ wave antennae. PhD thesis, Massachusetts Institute of Technology, 2004.
263
+ Behnke Berit. A Directed Search for Continuous Gravitational Waves from Unknown Iso-
264
+ lated Neutron Stars at the Galactic Center. PhD thesis, Leibniz University Hannover,
265
+ 2013.
266
+ Sylvia Biscoveanu, Maximiliano Isi, Salvatore Vitale, and Vijay Varma.
267
+ New Spin
268
+ on LIGO-Virgo Binary Black Holes.
269
+ Phys. Rev. Lett., 126(17):171103, 2021.
270
+ doi:
271
+ 10.1103/PhysRevLett.126.171103.
272
+ Lindy Blackburn. Open Issues in the Search for Gravitational Wave Transients. PhD thesis,
273
+ Massachusetts Institute of Technology, 2010.
274
+ Daniel E. Clark. Control of Differential Motion between Adjacent Advanced LIGO Seismic
275
+ Isolation Platforms. PhD thesis, Stanford University, 2013.
276
+ Katherine Laird Dooley. Design and performance of high laser power interferometers for
277
+ gravitational-wave detection. PhD thesis, University of Florida, 2011.
278
+ Jennifer Clair Driggers. Noise Cancellation for Gravitational Wave Detectors. PhD thesis,
279
+ California Institute of Technology, 2015.
280
+ Tobin Thomas Fricke. Homodyne detection for laser-interferometric gravitational wave de-
281
+ tectors. PhD thesis, Louisiana State University and Agricultural and Mechanical College,
282
+ 2011.
283
+ Paul D. Lasky et al. Gravitational-wave cosmology across 29 decades in frequency. Phys.
284
+ Rev. X, 6(1):011035, 2016. doi: 10.1103/PhysRevX.6.011035.
285
+ Michele Maggiore. Gravitational wave experiments and early universe cosmology. Phys.
286
+ Rept., 331:283–367, 2000. doi: 10.1016/S0370-1573(99)00102-7.
287
+ Denis Martynov. Lock Acquisition and Sensitivity Analysis of Advanced LIGO Interferom-
288
+ eters. PhD thesis, California Institute of Technology, 2015.
289
+ Ryan Quitzow-James. Search for Long-Duration Transient Gravitational Waves Associated
290
+ with Magnetar Bursts during LIGO’s Sixth Science Run. PhD thesis, Oregon U., 2016.
291
+ Joseph D. Romano and Neil J. Cornish. Detection methods for stochastic gravitational-
292
+ wave backgrounds: a unified treatment. Living Rev. Rel., 20(1):2, 2017. doi: 10.1007/
293
+ s41114-017-0004-1.
294
+ Michael P. Ross. Precision Mechanical Rotation Sensors for Terrestrial Gravitational Wave
295
+ Observatories. PhD thesis, University of Washington, 2020.
296
+ 6
297
+
298
+ Lightmorphic Signatures Analysis Toolkit
299
+ Darkhan Tuyenbayev. Extending the scientific reach of Advanced LIGO by compensating
300
+ for temporal variations in the calibration of the detectors. PhD thesis, The University of
301
+ Texas at San Antonio, 2017.
302
+ Madeline Wade. Gravitational-Wave Science with the Laser Interferometer Gravitational-
303
+ Wave Observatory. PhD thesis, University of Wisconsin–Milwaukee, 2015.
304
+ 7
305
+
49AyT4oBgHgl3EQfcPf_/content/tmp_files/load_file.txt ADDED
@@ -0,0 +1,160 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ filepath=/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf,len=159
2
+ page_content='arXiv:2301.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
3
+ page_content='00281v1 [cs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
4
+ page_content='LG] 31 Dec 2022 arXiv® 2022 (cs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
5
+ page_content='LG) 1-7 Submitted 12/22;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
6
+ page_content=' Published 12/22 Lightmorphic Signatures Analysis Toolkit Dumitru Damian dumitrudamian@yahoo.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
7
+ page_content='com Information and Communication Engineering Research and development consultant Timis,oara, RO Abstract In this paper we discuss the theory used in the design of an open source lightmorphic sig- natures analysis toolkit (LSAT).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
8
+ page_content=' In addition to providing a core functionality, the software package enables specific optimizations with its modular and customizable design.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
9
+ page_content=' To promote its usage and inspire future contributions, LSAT is publicly available.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
10
+ page_content=' By using a self-supervised neural network and augmented machine learning algorithms, LSAT provides an easy-to-use interface with ample documentation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
11
+ page_content=' The experiments demonstrate that LSAT improves the otherwise tedious and error- prone tasks of translating lightmorphic associated data into usable spectrograms, enhanced with parameter tuning and performance analysis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
12
+ page_content=' With the provided mathematical functions, LSAT validates the nonlinearity encoun- tered in the data conversion process while ensuring suitability of the forecasting algorithms.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
13
+ page_content=' Keywords: lightmorphic, machine learning, spectrogram, graph chord, neural network 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
14
+ page_content=' Introduction It is common knowledge, in the machine learning domain, to use differential values, since they provide a simple way to model the data.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
15
+ page_content=' However, such algorithms may not fit the lightmorphic signature properly, leading to a reduced quality of the obtained results.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
16
+ page_content=' Train- ing a neural network to predict the lightmorphic signature can significantly increase the data quality.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
17
+ page_content=' This is the task that LSAT tries to accomplish.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
18
+ page_content=' As such we define the lightmorphic metric learning (LML) as a branch of machine learning algorithms, set out with the purpose of learning lightmorphic signatures from multiple datasets trough usage of vibrating graph chords.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
19
+ page_content=' In the pursuing sections we describe the main features of the toolkit, explain the general mathematical concepts and finally detail the plans regarding future functionalities.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
20
+ page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
21
+ page_content=' General mathematical concepts In this section we expand the mathematical concepts and link them with the reasoning encountered in the implemented code.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
22
+ page_content=' We define the lightmorphic signature as a function of: light intensity (I) that varies according to seasons and local weather conditions, trajectory distribution characteristics ©2022 Dumitru Damian.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
23
+ page_content=' License: CC-BY 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
24
+ page_content='0, see https://creativecommons.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
25
+ page_content='org/licenses/by/4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
26
+ page_content='0/.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
27
+ page_content=' Typeset in LaTex using the JMLR LaTeX style file https://jmlr.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
28
+ page_content='org/author-info.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
29
+ page_content='html.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
30
+ page_content=' Damian (D), and specific adjustments (T): fL⊙ = I � 1 D � 1 T � 1 Γtζtdt (1) where: Γt – trajectory tensor ζt – point in time specificity Storage of these trajectory specific lightmorphic signatures is done in a database (Θ).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
31
+ page_content=' The segments containing isochronous surfaces with similarities are stored in another database (Φ) that serves as a baseline for training the neural network implementation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
32
+ page_content=' The isochronous surfaces that constitute the lightmorphic signature are interlinked trough the definition and usage of graph chords (δ(t)).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
33
+ page_content=' Observing their vibrational am- plitude allows the prediction of alternative lightmorphic signatures and, at the same time, correction of the already known values.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
34
+ page_content=' Since the primary light source considered is the Earth’s Sun, specific spacetime metrics (ex.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
35
+ page_content=' gµν, ηµν, h+, h×, Gµν) have to be used in order to describe the encountered anisotropies.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
36
+ page_content=' These are implemented as a function of distant astrophysical forces that stretch and com- press the fabric of spacetime.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
37
+ page_content=' According to special relativity, spacetime is seen as a four dimensional manifold de- scribed by a flat Minkowski metric defined in Cartesian coordinates (t, x, y, z, c = 1) as: ηµν = \uf8eb \uf8ec \uf8ec \uf8ed −1 0 0 0 0 1 0 0 0 0 1 0 0 0 0 1 \uf8f6 \uf8f7 \uf8f7 \uf8f8 , (2) When considering the geometry of curved space, we have made use of the metric gµν, that replaces the flat Minkowski metric ηµν.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
38
+ page_content=' This substitution was done considering that the geometry of curved space will eventually reduce to the flat spacetime of special relativity at a sufficiently small scale.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
39
+ page_content=' The interaction between curvature of spacetime and the mass distribution was modeled following (Blackburn (2010)) work, as: Gµν = kTµν (3) where Gµν is defined as the Einstein curvature tensor, Tµν is the stress-energy tensor and represents the mass-energy distribution, while k describes the Einstein constant of gravita- tion defined as: k = 8πG c4 (4) where c is the speed of light in a vacuum.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
40
+ page_content=' 2 Lightmorphic Signatures Analysis Toolkit At the same time, in order to improve the results quality, the Einstein tensor was also considered under the form: Gµν = Rµν − 1 2gµνR, (5) where Rµν is the Riemann tensor for the local spacetime, and R is the Ricci scalar.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
41
+ page_content=' Since there is not one general solution for the complex Einstein equations, but a large variety of possible solutions that apply to particular circumstances, we’ve considered a weak- field approximation, where the nonlinear Einstein equations where approximated towards linearity.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
42
+ page_content=' For example, a very small perturbation specific to a gravitational wave, will impact the flat spacetime and it is defined as hµν(x) and it’s value will be |hµν| << 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
43
+ page_content=' Thus, the Einstein equation becomes: gµν(x) = ηµν + hµν(x).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
44
+ page_content=' (6) or by simply considering the induced strain variations: □hµν(x) = 0, (7) By further pursuing such linearization, we can represent in the TT gauge, a propagating wave, under the following form: hTT µν = \uf8eb \uf8ec \uf8ec \uf8ed 0 0 0 0 0 h+ h× 0 0 h× −h+ 0 0 0 0 0 \uf8f6 \uf8f7 \uf8f7 \uf8f8 , (8) where the constant amplitudes (h+, h×) represent the two gravitational wave polariza- tions, the plus- and cross-polarization.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
45
+ page_content=' We represent the distance between two neighboring points as defined by (Berit (2013)) for a flat spacetime,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
46
+ page_content=' trough the following expression: ds2 = −c2dt2 + dx2 + dy2 + dz2 = −c2dt2 + [1 + h+(t)]dx2 + [1 − h+(t)]dy2 That allows us to model in the TT gauge,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
47
+ page_content=' the gravitational wave stretching along the x axis and compression along the y axis with the specific factor of: � 1 ± h+(t) ≃ 1 + 1 2h+(t) Having modeled the photon’s traveling path in outer space,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
48
+ page_content=' in order to simplify the inherent path inhomogeneities,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
49
+ page_content=' we separated the domains into outer space domain,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
50
+ page_content=' atmo- spheric domain and Earth specific domains (lithosphere,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
51
+ page_content=' hydrosphere,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
52
+ page_content=' biosphere,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
53
+ page_content=' noises,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
54
+ page_content=' etc).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
55
+ page_content=' We further define the phase of an electromagnetic wave of frequency ω0 as φ.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
56
+ page_content=' Following Driggers (2015)’s work, we consider that the starting light phase is at 0 and it travels at the speed of light c.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
57
+ page_content=' After a distance L it will have a phase δφspace that can be expressed as a distance integral over the spacetime metric, δφspace = ω0 c � L 0 gdx, (9) 3 Damian with g(t) = η+h(t), where η is the Minkowski metric and h(t) is the dimensionless spacetime strain.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
58
+ page_content=' Summing the light phase shift δφatm and the δφEarth which is derived from the noise sources like seismic or electromagnetic interferences,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
59
+ page_content=' leads to the dataset of trajectory spe- cific lightmorphic signatures: ΦΓIDT = N � j=1 Γj IDT (10) The signature parameter estimation is performed considering a prior distribution p(Φ|L⊙) that is updated upon receiving the new data d to give a posterior distribution p(Φ|d,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
60
+ page_content=' L⊙) p(Φ|d,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
61
+ page_content=' L⊙) = p(Φ|L⊙)p(d|Φ,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
62
+ page_content=' L⊙) p(d|L⊙) (11) While observing the distribution of multiple light segments within the dataset ΦΓIDT ,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
63
+ page_content=' it will be possible to estimate the probability for trajectory specific lightmorphic evolution: pΦ = f(ρk · pΦk) (12) where pΦk is the database’s k-th segment specific probability,' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
64
+ page_content=' ρk is the prediction weight for the k-th segment.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
65
+ page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
66
+ page_content=' Software package design The distribution matrices specific to the isochronous segmentation surfaces, which define the lightmorphic signature model, form the LSAT core.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
67
+ page_content=' As such we’ve used a design principle that ensures simplicity for the whole package, while making the source codes easy to read and maintain.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
68
+ page_content=' As the toolkit is written in a modular way, new functionalities can be easily plugged in.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
69
+ page_content=' This makes the LSAT not only a lightmorphic signature machine learning tool but also an experimental platform.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
70
+ page_content=' LSAT comes with plenty of documentation for all the interface functionalities and related data structures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
71
+ page_content=' The README file describes the installation process and interface usage.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
72
+ page_content=' For developers who use the toolkit in their applications, the API documentation can provide additional information related to functionality calls.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
73
+ page_content=' 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
74
+ page_content=' Practical Usage In the examples, we provide sample values for the lightmorphic signature updates, as a function of δφatm derived by the neural network from the values of a large dataset of at- mospheric meteorological data for 317 cities in Romania, with hundreds of thousands data points.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
75
+ page_content=' Automatic learning is supported trough API calls to the domain specific data providers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
76
+ page_content=' Beyond this simple way of running the lightmorphic signatures analysis toolkit, there are several enhancement options for advanced usage.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
77
+ page_content=' As example, one may activate additional functionalities that consider input parameters like complex space weather forecasting, dif- ferent electromagnetic wave disturbances or lithosphere, hydrosphere and biosphere specific localized data.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
78
+ page_content=' 4 Lightmorphic Signatures Analysis Toolkit 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
79
+ page_content=' Conclusion and Future Work With the lightmorphic signatures analysis toolkit we provided an open source SW package that is simple and easy-to-use.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
80
+ page_content=' Experiments and analysis conclude that the modular design and customization support are performing excellent in practice and can provide the base for additional research on lightmorphic signatures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
81
+ page_content=' The toolkit is constantly being improved by new research results and user feed-back with the ultimate goal of having an automated toolkit to use in maintaining and updating a large database of high-quality light signatures.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
82
+ page_content=' Future work will focus on probability estimates, additional functionalities that mitigate the large uncertainties in the available observational input data which arise from the complex interaction processes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
83
+ page_content=' In addition, the inclusion of artificial intelligence (AI) options will be considered while building a national/international network for lightmorphic signature analysis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
84
+ page_content=' 5 Damian 6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
85
+ page_content=' References References Rana Adhikari.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
86
+ page_content=' Sensitivity and noise analysis of 4 km laser interferometric gravitational wave antennae.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
87
+ page_content=' PhD thesis, Massachusetts Institute of Technology, 2004.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
88
+ page_content=' Behnke Berit.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
89
+ page_content=' A Directed Search for Continuous Gravitational Waves from Unknown Iso- lated Neutron Stars at the Galactic Center.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
90
+ page_content=' PhD thesis, Leibniz University Hannover, 2013.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
91
+ page_content=' Sylvia Biscoveanu, Maximiliano Isi, Salvatore Vitale, and Vijay Varma.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
92
+ page_content=' New Spin on LIGO-Virgo Binary Black Holes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
93
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
94
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
95
+ page_content=' Lett.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
96
+ page_content=', 126(17):171103, 2021.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
97
+ page_content=' doi: 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
98
+ page_content='1103/PhysRevLett.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
99
+ page_content='126.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
100
+ page_content='171103.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
101
+ page_content=' Lindy Blackburn.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
102
+ page_content=' Open Issues in the Search for Gravitational Wave Transients.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
103
+ page_content=' PhD thesis, Massachusetts Institute of Technology, 2010.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
104
+ page_content=' Daniel E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
105
+ page_content=' Clark.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
106
+ page_content=' Control of Differential Motion between Adjacent Advanced LIGO Seismic Isolation Platforms.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
107
+ page_content=' PhD thesis, Stanford University, 2013.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
108
+ page_content=' Katherine Laird Dooley.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
109
+ page_content=' Design and performance of high laser power interferometers for gravitational-wave detection.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
110
+ page_content=' PhD thesis, University of Florida, 2011.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
111
+ page_content=' Jennifer Clair Driggers.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
112
+ page_content=' Noise Cancellation for Gravitational Wave Detectors.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
113
+ page_content=' PhD thesis, California Institute of Technology, 2015.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
114
+ page_content=' Tobin Thomas Fricke.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
115
+ page_content=' Homodyne detection for laser-interferometric gravitational wave de- tectors.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
116
+ page_content=' PhD thesis, Louisiana State University and Agricultural and Mechanical College, 2011.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
117
+ page_content=' Paul D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
118
+ page_content=' Lasky et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
119
+ page_content=' Gravitational-wave cosmology across 29 decades in frequency.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
120
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
121
+ page_content=' Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
122
+ page_content=' X, 6(1):011035, 2016.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
123
+ page_content=' doi: 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
124
+ page_content='1103/PhysRevX.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
125
+ page_content='6.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
126
+ page_content='011035.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
127
+ page_content=' Michele Maggiore.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
128
+ page_content=' Gravitational wave experiments and early universe cosmology.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
129
+ page_content=' Phys.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
130
+ page_content=' Rept.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
131
+ page_content=', 331:283–367, 2000.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
132
+ page_content=' doi: 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
133
+ page_content='1016/S0370-1573(99)00102-7.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
134
+ page_content=' Denis Martynov.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
135
+ page_content=' Lock Acquisition and Sensitivity Analysis of Advanced LIGO Interferom- eters.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
136
+ page_content=' PhD thesis, California Institute of Technology, 2015.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
137
+ page_content=' Ryan Quitzow-James.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
138
+ page_content=' Search for Long-Duration Transient Gravitational Waves Associated with Magnetar Bursts during LIGO’s Sixth Science Run.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
139
+ page_content=' PhD thesis, Oregon U.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
140
+ page_content=', 2016.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
141
+ page_content=' Joseph D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
142
+ page_content=' Romano and Neil J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
143
+ page_content=' Cornish.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
144
+ page_content=' Detection methods for stochastic gravitational- wave backgrounds: a unified treatment.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
145
+ page_content=' Living Rev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
146
+ page_content=' Rel.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
147
+ page_content=', 20(1):2, 2017.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
148
+ page_content=' doi: 10.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
149
+ page_content='1007/ s41114-017-0004-1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
150
+ page_content=' Michael P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
151
+ page_content=' Ross.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
152
+ page_content=' Precision Mechanical Rotation Sensors for Terrestrial Gravitational Wave Observatories.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
153
+ page_content=' PhD thesis, University of Washington, 2020.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
154
+ page_content=' 6 Lightmorphic Signatures Analysis Toolkit Darkhan Tuyenbayev.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
155
+ page_content=' Extending the scientific reach of Advanced LIGO by compensating for temporal variations in the calibration of the detectors.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
156
+ page_content=' PhD thesis, The University of Texas at San Antonio, 2017.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
157
+ page_content=' Madeline Wade.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
158
+ page_content=' Gravitational-Wave Science with the Laser Interferometer Gravitational- Wave Observatory.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
159
+ page_content=' PhD thesis, University of Wisconsin–Milwaukee, 2015.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
160
+ page_content=' 7' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/49AyT4oBgHgl3EQfcPf_/content/2301.00281v1.pdf'}
4NFAT4oBgHgl3EQfExwU/content/tmp_files/2301.08423v1.pdf.txt ADDED
The diff for this file is too large to render. See raw diff
 
4NFAT4oBgHgl3EQfExwU/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
59E3T4oBgHgl3EQfQwmQ/content/tmp_files/2301.04416v1.pdf.txt ADDED
@@ -0,0 +1,237 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ pyssam – a Python library for statistical modelling of
2
+ biomedical shape and appearance
3
+ Josh Williams1, Ali Ozel1, and Uwe Wolfram1
4
+ 1 School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh, UK
5
+ DOI: TBD
6
+ Software
7
+ • Repository
8
+ • Archive
9
+ Editor: Pending Editor
10
+ Reviewers:
11
+ • Pending Reviewers
12
+ Submitted: N/A
13
+ Published: N/A
14
+ License
15
+ Authors of papers retain
16
+ copyright and release the work
17
+ under a Creative Commons
18
+ Attribution 4.0 International
19
+ License (CC BY 4.0).
20
+ Summary
21
+ pyssam is a Python library for creating statistical shape and appearance models (SSAMs)
22
+ for biological (and other) shapes such as bones, lungs or other organs. A point cloud best
23
+ describing the anatomical ‘landmarks’ of the organ are required from each sample in a small
24
+ population as an input. Additional information such as landmark gray-value can be included to
25
+ incorporate joint correlations of shape and ‘appearance’ into the model. Our library performs
26
+ alignment and scaling of the input data and creates a SSAM based on covariance across the
27
+ population. The output SSAM can be used to parameterise and quantify shape change across
28
+ a population. pyssam is a small and low dependency codebase with examples included as
29
+ Jupyter notebooks for several common SSAM computations. The given examples can easily be
30
+ extended to alternative datasets, and also alternative tasks such as medical image segmentation
31
+ by incorporating a SSAM as a constraint for segmented organs.
32
+ Statement of need
33
+ Statistical shape (and appearance) models (SSAMs) have drawn significant interest in biomed-
34
+ ical engineering and computer vision research due to their ability to automatically deduce a
35
+ linear parameterisation of shape covariances across a small population of training data (Baka
36
+ et al., 2011; Cootes et al., 1995; Heimann & Meinzer, 2009; Väänänen et al., 2015). The
37
+ classic statistical shape model (SSM) approach uses a point cloud of landmarks which are
38
+ in correspondence across several instances of a shape. The covariances of how the shape
39
+ changes across the training population are computed, and principal component analysis (PCA)
40
+ is used to parameterise the different modes of shape variation (Cootes et al., 1995). This
41
+ approach paved the way for automatic algorithms which could significantly aid medical image
42
+ segmentation (similar to an atlas) (Irving et al., 2011), characterise how the organ shape varies
43
+ over a population as a diagnostic tool (Osanlouy et al., 2020), or even reconstruct a full 3D
44
+ structure from a sparser imaging modality such as planar X-ray images (Baka et al., 2011;
45
+ Väänänen et al., 2015).
46
+ We have found that available open-source toolkits such as Statismo and Scalismo (Lüthi et
47
+ al., 2012) suffer from an exhaustive number of dependencies and are difficult to adapt to
48
+ new tasks, datasets and I/O datatypes. ShapeWorks (Cates et al., 2017) is another strongly
49
+ developed library for statistical shape modelling, but it uses an alternative method of extracting
50
+ landmarks (a so-called particle-based method) which is less broadly used and more complex
51
+ than a landmark-based system (where landmarks can be defined in any desired way for different
52
+ anatomical shapes). Additionally, as the machine learning ecosystem has strong foundations in
53
+ Python, building statistical models in C++, Scala or other languages reduces compatibility
54
+ with the majority of modern machine learning developments (Bhalodia et al., 2018). We
55
+ therefore implemented a lightweight Python framework for SSAMs which is easily adaptable
56
+ with few dependencies, making it suitable for integrating as part of a broader codebase, as
57
+ well as installing and running on high-performance computing clusters where users do not have
58
+ root access to install many dependencies. We provide Jupyter notebooks on readthedocs and
59
+ Williams et al. (2023). pyssam – a Python library for statistical modelling of biomedical shape and appearance. Journal of Open Source Software,
60
+ TBD, TBD. https://doi.org/TBD
61
+ 1
62
+ arXiv:2301.04416v1 [q-bio.QM] 11 Jan 2023
63
+
64
+ two example datasets that allow users new to coding or SSAMs to learn how these models
65
+ work in an interactive way to ease access when learning a new research topic and library.
66
+ Overview
67
+ The main modelling classes are built on the abstract base class StatisticalModelBase,
68
+ which has several methods for pre-processing data and performing PCA (Figure 1). There
69
+ are also several global variables that are inherited which are related to principal components,
70
+ component variances and model parameters. The classes for SSM and SAM pre-process the data
71
+ (align to zero mean and standard deviation of one) and can compute the population mean
72
+ shape/appearance. Finally, the SSAM class for shape and appearance modelling inherits all
73
+ of these, but also imports the SSM and SAM methods to pre-process shape and appearance
74
+ features separately, before they are merged into one dataset for modelling.
75
+ StatisticalModelBase
76
+ SSM
77
+ StatisticalModelBase
78
+ SAM
79
+ SSAM
80
+ StatisticalModelBase
81
+ StatisticalModelBase
82
+ SSM
83
+ SAM
84
+ Figure 1: Schematic overview of the codebase. Each modelling class is abstracted from the Statis
85
+ ticalModelBase class and contains several inherited variables such as model weights and principal
86
+ components. The SSAM class inherits from StatisticalModelBase, but also uses pre-processing
87
+ pipelines from SSM and SAM.
88
+ Examples
89
+ Here we present two example applications of pyssam. The first example examines shape
90
+ variations in a toy dataset created for this study, which has a tree structure. Tree structures
91
+ appear often in biology, including the lung airways and vascular system. Toy datasets such as
92
+ these are a simple means to visualise and interpret the modelling and code framework. We then
93
+ provide a more complex example which considers the left lower lobe of human lungs obtained
94
+ from CT data (Tang et al., 2019). This example considers shape and appearance, where the
95
+ appearance is the gray-value at the landmark location on an X-ray projection (obtained with
96
+ the AppearanceFromXray helper class).
97
+ Statistical shape modelling toy dataset
98
+ To understand the shape modelling process, we have provided a dataset class called Tree
99
+ which creates a number of tree shapes which are randomly computed based on global minimum
100
+ and maximum values for angle and branch length ratio (between parent and child). Tree
101
+ parameters are shown in Figure 2a. Tree nodes are converted to a numpy array and used to
102
+ initialise pyssam.SSM. At initialisation of the SSM class, the landmarks are aligned, scaled to
103
+ unit standard deviation and stacked into a matrix of shape (Nf, 3NL) where Nf is the number
104
+ of features (samples in our training dataset) and NL is the number of landmarks (each with
105
+ a x, y, z coordinates). All y coordinates in this case are zero, meaning the data is actually
106
+ 2D but we preserve a 3D coordinate system for simplicity in generalising the code to more
107
+ Williams et al. (2023). pyssam – a Python library for statistical modelling of biomedical shape and appearance. Journal of Open Source Software,
108
+ TBD, TBD. https://doi.org/TBD
109
+ 2
110
+
111
+ common 3D applications. The code below shows how we can simply obtain a SSM from a set
112
+ of landmarks.
113
+ from glob import glob
114
+ import numpy as np
115
+ import pyssam
116
+ tree_class = pyssam.datasets.Tree(num_extra_ends=1)
117
+ landmark_coordinates = np.array(
118
+ [tree_class.make_tree_landmarks() for i in range(0, num_samples)]
119
+ )
120
+ ssm_obj = pyssam.SSM(landmark_coordinates)
121
+ ssm_obj.create_pca_model(ssm_obj.landmarks_scale)
122
+ mean_shape_columnvector = ssm_obj.compute_dataset_mean()
123
+ L1
124
+ L2
125
+ θ
126
+ 0
127
+ 10
128
+ 20
129
+ 30
130
+ 40
131
+ Number of components
132
+ 50
133
+ 60
134
+ 70
135
+ 80
136
+ 90
137
+ 100
138
+ Variance [%]
139
+ (a)
140
+ (b)
141
+ Figure 2: Overview of tree dataset population. Panels show (a) a visualisation of 100 tree samples,
142
+ and (b) cumulative variance versus the number of PCA components constructed by the statistical
143
+ shape model. Inset of (a) shows a legend describing the morphological parameters varied to create
144
+ the tree dataset. These parameters include the initial branch length, L1, the branch length ratio
145
+ LR = L2/L1, and branching angle θ.
146
+ Shape and appearance modelling of lung shape and chest X-ray images
147
+ In the following example, we show a real application where 3D landmark for the left lower
148
+ lung lobe are projected onto digitally reconstructed X-rays (Väänänen et al., 2015) and the
149
+ gray-value is used to obtain appearance. Example landmark data was obtained using an
150
+ automatic algorithm (Ferrarini et al., 2007). Appearance information is extracted from the
151
+ X-ray images using AppearanceFromXray (part of pyssam.utils). We use landmarks,
152
+ X-ray images as well as origin and pixel spacing information for the X-ray images to extract
153
+ appearance as follows
154
+ appearance_xr = pyssam.AppearanceFromXray(
155
+ IMAGE_DATASET, IMAGE_ORIGIN, IMAGE_SPACING
156
+ )
157
+ appearance_values = appearance_xr.all_landmark_density(
158
+ landmarks_coordinates
159
+ )
160
+ Williams et al. (2023). pyssam – a Python library for statistical modelling of biomedical shape and appearance. Journal of Open Source Software,
161
+ TBD, TBD. https://doi.org/TBD
162
+ 3
163
+
164
+ The SSAM can then be trained in a similar way as the SSM in subsection with the following
165
+ code snippet:
166
+ ssam_obj = pyssam.SSAM(landmark_coordinates, appearance_values)
167
+ ssam_obj.create_pca_model(ssam_obj.shape_appearance_columns)
168
+ mean_shape_appearance_columnvector = ssam_obj.compute_dataset_mean()
169
+ The shape and appearance modes can then be computed based on the model parameters
170
+ (ssam.model_parameters). The computed model parameters (eigenvectors and eigenvalues
171
+ of the covariance matrix) can be used to morph the shape and appearance using ssam.morph
172
+ _model (part of StatisticalModelBase in Figure 1) by
173
+ x ≈ ¯x + Φ · b
174
+ (1)
175
+ where x is a new array containing shape and appearance, ¯x is the training dataset mean
176
+ shape and appearance, Φ is the model principal components (eigenvectors of the training data
177
+ covariance matrix), b is the model parameters, which is an array of weights unique to each
178
+ data sample. The model parameter a mode m should be within [−3
179
+
180
+ σ2
181
+ m, 3
182
+
183
+ σ2
184
+ m], where
185
+ σ2
186
+ m is the explained variance of m (mth largest eigenvalue of the covariance matrix) (Cootes
187
+ et al., 1995).
188
+ Each mode of shape and appearance variation is visualised, as shown for a representative mode
189
+ in Figure 3. This shows how lung shape influences the gray-value of lung pixels on the X-ray
190
+ image. In this case, the change in shape and appearance are mainly due to how the lung
191
+ interacts with adjacent structures such as the heart, rib cage and diaphragm.
192
+ Figure 3: First mode of SSAM variation for lung lobe dataset. Panels show shape and appearance
193
+ morphed using ssam.morph_model method and varying the model parameters (ssam.model_parame
194
+ ters), from -2, 0 (mean shape) and 2.
195
+ Acknowledgement
196
+ JW was funded by a 2019 PhD Scholarship from the Carnegie-Trust for the Universities of
197
+ Scotland.
198
+ Williams et al. (2023). pyssam – a Python library for statistical modelling of biomedical shape and appearance. Journal of Open Source Software,
199
+ TBD, TBD. https://doi.org/TBD
200
+ 4
201
+
202
+ References
203
+ Baka, N., Kaptein, B. L., Bruijne, M. de, Walsum, T. van, Giphart, J., Niessen, W. J., &
204
+ Lelieveldt, B. P. (2011). 2D–3D shape reconstruction of the distal femur from stereo X-ray
205
+ imaging using statistical shape models. Medical Image Analysis, 15(6), 840–850.
206
+ Bhalodia, R., Elhabian, S. Y., Kavan, L., & Whitaker, R. T. (2018). DeepSSM: A deep
207
+ learning framework for statistical shape modeling from raw images. International Workshop
208
+ on Shape in Medical Imaging, 244–257.
209
+ Cates, J., Elhabian, S., & Whitaker, R. (2017). Shapeworks: Particle-based shape corre-
210
+ spondence and visualization software. In Statistical shape and deformation analysis (pp.
211
+ 257–298). Elsevier.
212
+ Cootes, T. F., Taylor, C. J., Cooper, D. H., & Graham, J. (1995). Active shape models-their
213
+ training and application. Computer Vision and Image Understanding, 61(1), 38–59.
214
+ Ferrarini, L., Olofsen, H., Palm, W. M., Van Buchem, M. A., Reiber, J. H., & Admiraal-Behloul,
215
+ F. (2007). GAMEs: Growing and adaptive meshes for fully automatic shape modeling and
216
+ analysis. Medical Image Analysis, 11(3), 302–314.
217
+ Heimann, T., & Meinzer, H.-P. (2009).
218
+ Statistical shape models for 3D medical image
219
+ segmentation: A review. Medical Image Analysis, 13(4), 543–563.
220
+ Irving, B., Goussard, P., Gie, R., Todd-Pokropek, A., & Taylor, P. (2011). Segmentation of
221
+ obstructed airway branches in CT using airway topology and statistical shape analysis. 2011
222
+ IEEE International Symposium on Biomedical Imaging: From Nano to Macro, 447–451.
223
+ Lüthi, M., Blanc, R., Albrecht, T., Gass, T., Goksel, O., Büchler, P., Kistler, M., Bousleiman,
224
+ H., Reyes, M., Cattin, P., & others. (2012). Statismo-a framework for PCA based statistical
225
+ models. The Insight Journal, 2012, 1–18.
226
+ Osanlouy, M., Clark, A. R., Kumar, H., King, C., Wilsher, M. L., Milne, D. G., Whyte, K.,
227
+ Hoffman, E. A., & Tawhai, M. H. (2020). Lung and fissure shape is associated with age in
228
+ healthy never-smoking adults aged 20–90 years. Scientific Reports, 10(1), 1–13.
229
+ Tang, H., Zhang, C., & Xie, X. (2019). Automatic pulmonary lobe segmentation using deep
230
+ learning. arXiv Preprint arXiv:1903.09879.
231
+ Väänänen, S. P., Grassi, L., Flivik, G., Jurvelin, J. S., & Isaksson, H. (2015). Generation of
232
+ 3D shape, density, cortical thickness and finite element mesh of proximal femur from a
233
+ DXA image. Medical Image Analysis, 24(1), 125–134.
234
+ Williams et al. (2023). pyssam – a Python library for statistical modelling of biomedical shape and appearance. Journal of Open Source Software,
235
+ TBD, TBD. https://doi.org/TBD
236
+ 5
237
+
59E3T4oBgHgl3EQfQwmQ/content/tmp_files/load_file.txt ADDED
@@ -0,0 +1,236 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ filepath=/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf,len=235
2
+ page_content='pyssam – a Python library for statistical modelling of biomedical shape and appearance Josh Williams1, Ali Ozel1, and Uwe Wolfram1 1 School of Engineering and Physical Sciences, Heriot-Watt University, Edinburgh, UK DOI: TBD Software Repository Archive Editor: Pending Editor Reviewers: Pending Reviewers Submitted: N/A Published: N/A License Authors of papers retain copyright and release the work under a Creative Commons Attribution 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
3
+ page_content='0 International License (CC BY 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
4
+ page_content='0).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
5
+ page_content=' Summary pyssam is a Python library for creating statistical shape and appearance models (SSAMs) for biological (and other) shapes such as bones, lungs or other organs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
6
+ page_content=' A point cloud best describing the anatomical ‘landmarks’ of the organ are required from each sample in a small population as an input.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
7
+ page_content=' Additional information such as landmark gray-value can be included to incorporate joint correlations of shape and ‘appearance’ into the model.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
8
+ page_content=' Our library performs alignment and scaling of the input data and creates a SSAM based on covariance across the population.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
9
+ page_content=' The output SSAM can be used to parameterise and quantify shape change across a population.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
10
+ page_content=' pyssam is a small and low dependency codebase with examples included as Jupyter notebooks for several common SSAM computations.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
11
+ page_content=' The given examples can easily be extended to alternative datasets, and also alternative tasks such as medical image segmentation by incorporating a SSAM as a constraint for segmented organs.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
12
+ page_content=' Statement of need Statistical shape (and appearance) models (SSAMs) have drawn significant interest in biomed- ical engineering and computer vision research due to their ability to automatically deduce a linear parameterisation of shape covariances across a small population of training data (Baka et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
13
+ page_content=', 2011;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
14
+ page_content=' Cootes et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
15
+ page_content=', 1995;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
16
+ page_content=' Heimann & Meinzer, 2009;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
17
+ page_content=' Väänänen et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
18
+ page_content=', 2015).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
19
+ page_content=' The classic statistical shape model (SSM) approach uses a point cloud of landmarks which are in correspondence across several instances of a shape.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
20
+ page_content=' The covariances of how the shape changes across the training population are computed, and principal component analysis (PCA) is used to parameterise the different modes of shape variation (Cootes et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
21
+ page_content=', 1995).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
22
+ page_content=' This approach paved the way for automatic algorithms which could significantly aid medical image segmentation (similar to an atlas) (Irving et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
23
+ page_content=', 2011), characterise how the organ shape varies over a population as a diagnostic tool (Osanlouy et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
24
+ page_content=', 2020), or even reconstruct a full 3D structure from a sparser imaging modality such as planar X-ray images (Baka et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
25
+ page_content=', 2011;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
26
+ page_content=' Väänänen et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
27
+ page_content=', 2015).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
28
+ page_content=' We have found that available open-source toolkits such as Statismo and Scalismo (Lüthi et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
29
+ page_content=', 2012) suffer from an exhaustive number of dependencies and are difficult to adapt to new tasks, datasets and I/O datatypes.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
30
+ page_content=' ShapeWorks (Cates et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
31
+ page_content=', 2017) is another strongly developed library for statistical shape modelling, but it uses an alternative method of extracting landmarks (a so-called particle-based method) which is less broadly used and more complex than a landmark-based system (where landmarks can be defined in any desired way for different anatomical shapes).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
32
+ page_content=' Additionally, as the machine learning ecosystem has strong foundations in Python, building statistical models in C++, Scala or other languages reduces compatibility with the majority of modern machine learning developments (Bhalodia et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
33
+ page_content=', 2018).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
34
+ page_content=' We therefore implemented a lightweight Python framework for SSAMs which is easily adaptable with few dependencies, making it suitable for integrating as part of a broader codebase, as well as installing and running on high-performance computing clusters where users do not have root access to install many dependencies.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
35
+ page_content=' We provide Jupyter notebooks on readthedocs and Williams et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
36
+ page_content=' (2023).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
37
+ page_content=' pyssam – a Python library for statistical modelling of biomedical shape and appearance.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
38
+ page_content=' Journal of Open Source Software, TBD, TBD.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
39
+ page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
40
+ page_content='org/TBD 1 arXiv:2301.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
41
+ page_content='04416v1 [q-bio.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
42
+ page_content='QM] 11 Jan 2023 two example datasets that allow users new to coding or SSAMs to learn how these models work in an interactive way to ease access when learning a new research topic and library.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
43
+ page_content=' Overview The main modelling classes are built on the abstract base class StatisticalModelBase, which has several methods for pre-processing data and performing PCA (Figure 1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
44
+ page_content=' There are also several global variables that are inherited which are related to principal components, component variances and model parameters.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
45
+ page_content=' The classes for SSM and SAM pre-process the data (align to zero mean and standard deviation of one) and can compute the population mean shape/appearance.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
46
+ page_content=' Finally, the SSAM class for shape and appearance modelling inherits all of these, but also imports the SSM and SAM methods to pre-process shape and appearance features separately, before they are merged into one dataset for modelling.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
47
+ page_content=' StatisticalModelBase SSM StatisticalModelBase SAM SSAM StatisticalModelBase StatisticalModelBase SSM SAM Figure 1: Schematic overview of the codebase.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
48
+ page_content=' Each modelling class is abstracted from the Statis ticalModelBase class and contains several inherited variables such as model weights and principal components.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
49
+ page_content=' The SSAM class inherits from StatisticalModelBase, but also uses pre-processing pipelines from SSM and SAM.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
50
+ page_content=' Examples Here we present two example applications of pyssam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
51
+ page_content=' The first example examines shape variations in a toy dataset created for this study, which has a tree structure.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
52
+ page_content=' Tree structures appear often in biology, including the lung airways and vascular system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
53
+ page_content=' Toy datasets such as these are a simple means to visualise and interpret the modelling and code framework.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
54
+ page_content=' We then provide a more complex example which considers the left lower lobe of human lungs obtained from CT data (Tang et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
55
+ page_content=', 2019).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
56
+ page_content=' This example considers shape and appearance, where the appearance is the gray-value at the landmark location on an X-ray projection (obtained with the AppearanceFromXray helper class).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
57
+ page_content=' Statistical shape modelling toy dataset To understand the shape modelling process, we have provided a dataset class called Tree which creates a number of tree shapes which are randomly computed based on global minimum and maximum values for angle and branch length ratio (between parent and child).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
58
+ page_content=' Tree parameters are shown in Figure 2a.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
59
+ page_content=' Tree nodes are converted to a numpy array and used to initialise pyssam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
60
+ page_content='SSM.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
61
+ page_content=' At initialisation of the SSM class, the landmarks are aligned, scaled to unit standard deviation and stacked into a matrix of shape (Nf, 3NL) where Nf is the number of features (samples in our training dataset) and NL is the number of landmarks (each with a x, y, z coordinates).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
62
+ page_content=' All y coordinates in this case are zero, meaning the data is actually 2D but we preserve a 3D coordinate system for simplicity in generalising the code to more Williams et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
63
+ page_content=' (2023).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
64
+ page_content=' pyssam – a Python library for statistical modelling of biomedical shape and appearance.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
65
+ page_content=' Journal of Open Source Software, TBD, TBD.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
66
+ page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
67
+ page_content='org/TBD 2 common 3D applications.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
68
+ page_content=' The code below shows how we can simply obtain a SSM from a set of landmarks.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
69
+ page_content=' from glob import glob import numpy as np import pyssam tree_class = pyssam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
70
+ page_content='datasets.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
71
+ page_content='Tree(num_extra_ends=1) landmark_coordinates = np.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
72
+ page_content='array( [tree_class.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
73
+ page_content='make_tree_landmarks() for i in range(0, num_samples)] ) ssm_obj = pyssam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
74
+ page_content='SSM(landmark_coordinates) ssm_obj.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
75
+ page_content='create_pca_model(ssm_obj.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
76
+ page_content='landmarks_scale) mean_shape_columnvector = ssm_obj.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
77
+ page_content='compute_dataset_mean() L1 L2 θ 0 10 20 30 40 Number of components 50 60 70 80 90 100 Variance [%] (a) (b) Figure 2: Overview of tree dataset population.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
78
+ page_content=' Panels show (a) a visualisation of 100 tree samples, and (b) cumulative variance versus the number of PCA components constructed by the statistical shape model.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
79
+ page_content=' Inset of (a) shows a legend describing the morphological parameters varied to create the tree dataset.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
80
+ page_content=' These parameters include the initial branch length, L1, the branch length ratio LR = L2/L1, and branching angle θ.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
81
+ page_content=' Shape and appearance modelling of lung shape and chest X-ray images In the following example, we show a real application where 3D landmark for the left lower lung lobe are projected onto digitally reconstructed X-rays (Väänänen et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
82
+ page_content=', 2015) and the gray-value is used to obtain appearance.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
83
+ page_content=' Example landmark data was obtained using an automatic algorithm (Ferrarini et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
84
+ page_content=', 2007).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
85
+ page_content=' Appearance information is extracted from the X-ray images using AppearanceFromXray (part of pyssam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
86
+ page_content='utils).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
87
+ page_content=' We use landmarks, X-ray images as well as origin and pixel spacing information for the X-ray images to extract appearance as follows appearance_xr = pyssam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
88
+ page_content='AppearanceFromXray( IMAGE_DATASET, IMAGE_ORIGIN, IMAGE_SPACING ) appearance_values = appearance_xr.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
89
+ page_content='all_landmark_density( landmarks_coordinates ) Williams et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
90
+ page_content=' (2023).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
91
+ page_content=' pyssam – a Python library for statistical modelling of biomedical shape and appearance.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
92
+ page_content=' Journal of Open Source Software, TBD, TBD.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
93
+ page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
94
+ page_content='org/TBD 3 The SSAM can then be trained in a similar way as the SSM in subsection with the following code snippet: ssam_obj = pyssam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
95
+ page_content='SSAM(landmark_coordinates, appearance_values) ssam_obj.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
96
+ page_content='create_pca_model(ssam_obj.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
97
+ page_content='shape_appearance_columns) mean_shape_appearance_columnvector = ssam_obj.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
98
+ page_content='compute_dataset_mean() The shape and appearance modes can then be computed based on the model parameters (ssam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
99
+ page_content='model_parameters).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
100
+ page_content=' The computed model parameters (eigenvectors and eigenvalues of the covariance matrix) can be used to morph the shape and appearance using ssam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
101
+ page_content='morph _model (part of StatisticalModelBase in Figure 1) by x ≈ ¯x + Φ · b (1) where x is a new array containing shape and appearance, ¯x is the training dataset mean shape and appearance, Φ is the model principal components (eigenvectors of the training data covariance matrix), b is the model parameters, which is an array of weights unique to each data sample.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
102
+ page_content=' The model parameter a mode m should be within [−3 � σ2 m, 3 � σ2 m], where σ2 m is the explained variance of m (mth largest eigenvalue of the covariance matrix) (Cootes et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
103
+ page_content=', 1995).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
104
+ page_content=' Each mode of shape and appearance variation is visualised, as shown for a representative mode in Figure 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
105
+ page_content=' This shows how lung shape influences the gray-value of lung pixels on the X-ray image.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
106
+ page_content=' In this case, the change in shape and appearance are mainly due to how the lung interacts with adjacent structures such as the heart, rib cage and diaphragm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
107
+ page_content=' Figure 3: First mode of SSAM variation for lung lobe dataset.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
108
+ page_content=' Panels show shape and appearance morphed using ssam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
109
+ page_content='morph_model method and varying the model parameters (ssam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
110
+ page_content='model_parame ters), from -2, 0 (mean shape) and 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
111
+ page_content=' Acknowledgement JW was funded by a 2019 PhD Scholarship from the Carnegie-Trust for the Universities of Scotland.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
112
+ page_content=' Williams et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
113
+ page_content=' (2023).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
114
+ page_content=' pyssam – a Python library for statistical modelling of biomedical shape and appearance.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
115
+ page_content=' Journal of Open Source Software, TBD, TBD.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
116
+ page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
117
+ page_content='org/TBD 4 References Baka, N.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
118
+ page_content=', Kaptein, B.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
119
+ page_content=' L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
120
+ page_content=', Bruijne, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
121
+ page_content=' de, Walsum, T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
122
+ page_content=' van, Giphart, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
123
+ page_content=', Niessen, W.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
124
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
125
+ page_content=', & Lelieveldt, B.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
126
+ page_content=' P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
127
+ page_content=' (2011).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
128
+ page_content=' 2D–3D shape reconstruction of the distal femur from stereo X-ray imaging using statistical shape models.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
129
+ page_content=' Medical Image Analysis, 15(6), 840–850.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
130
+ page_content=' Bhalodia, R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
131
+ page_content=', Elhabian, S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
132
+ page_content=' Y.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
133
+ page_content=', Kavan, L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
134
+ page_content=', & Whitaker, R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
135
+ page_content=' T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
136
+ page_content=' (2018).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
137
+ page_content=' DeepSSM: A deep learning framework for statistical shape modeling from raw images.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
138
+ page_content=' International Workshop on Shape in Medical Imaging, 244–257.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
139
+ page_content=' Cates, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
140
+ page_content=', Elhabian, S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
141
+ page_content=', & Whitaker, R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
142
+ page_content=' (2017).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
143
+ page_content=' Shapeworks: Particle-based shape corre- spondence and visualization software.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
144
+ page_content=' In Statistical shape and deformation analysis (pp.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
145
+ page_content=' 257–298).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
146
+ page_content=' Elsevier.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
147
+ page_content=' Cootes, T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
148
+ page_content=' F.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
149
+ page_content=', Taylor, C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
150
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
151
+ page_content=', Cooper, D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
152
+ page_content=' H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
153
+ page_content=', & Graham, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
154
+ page_content=' (1995).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
155
+ page_content=' Active shape models-their training and application.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
156
+ page_content=' Computer Vision and Image Understanding, 61(1), 38–59.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
157
+ page_content=' Ferrarini, L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
158
+ page_content=', Olofsen, H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
159
+ page_content=', Palm, W.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
160
+ page_content=' M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
161
+ page_content=', Van Buchem, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
162
+ page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
163
+ page_content=', Reiber, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
164
+ page_content=' H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
165
+ page_content=', & Admiraal-Behloul, F.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
166
+ page_content=' (2007).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
167
+ page_content=' GAMEs: Growing and adaptive meshes for fully automatic shape modeling and analysis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
168
+ page_content=' Medical Image Analysis, 11(3), 302–314.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
169
+ page_content=' Heimann, T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
170
+ page_content=', & Meinzer, H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
171
+ page_content='-P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
172
+ page_content=' (2009).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
173
+ page_content=' Statistical shape models for 3D medical image segmentation: A review.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
174
+ page_content=' Medical Image Analysis, 13(4), 543–563.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
175
+ page_content=' Irving, B.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
176
+ page_content=', Goussard, P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
177
+ page_content=', Gie, R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
178
+ page_content=', Todd-Pokropek, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
179
+ page_content=', & Taylor, P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
180
+ page_content=' (2011).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
181
+ page_content=' Segmentation of obstructed airway branches in CT using airway topology and statistical shape analysis.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
182
+ page_content=' 2011 IEEE International Symposium on Biomedical Imaging: From Nano to Macro, 447–451.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
183
+ page_content=' Lüthi, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
184
+ page_content=', Blanc, R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
185
+ page_content=', Albrecht, T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
186
+ page_content=', Gass, T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
187
+ page_content=', Goksel, O.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
188
+ page_content=', Büchler, P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
189
+ page_content=', Kistler, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
190
+ page_content=', Bousleiman, H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
191
+ page_content=', Reyes, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
192
+ page_content=', Cattin, P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
193
+ page_content=', & others.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
194
+ page_content=' (2012).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
195
+ page_content=' Statismo-a framework for PCA based statistical models.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
196
+ page_content=' The Insight Journal, 2012, 1–18.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
197
+ page_content=' Osanlouy, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
198
+ page_content=', Clark, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
199
+ page_content=' R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
200
+ page_content=', Kumar, H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
201
+ page_content=', King, C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
202
+ page_content=', Wilsher, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
203
+ page_content=' L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
204
+ page_content=', Milne, D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
205
+ page_content=' G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
206
+ page_content=', Whyte, K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
207
+ page_content=', Hoffman, E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
208
+ page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
209
+ page_content=', & Tawhai, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
210
+ page_content=' H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
211
+ page_content=' (2020).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
212
+ page_content=' Lung and fissure shape is associated with age in healthy never-smoking adults aged 20–90 years.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
213
+ page_content=' Scientific Reports, 10(1), 1–13.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
214
+ page_content=' Tang, H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
215
+ page_content=', Zhang, C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
216
+ page_content=', & Xie, X.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
217
+ page_content=' (2019).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
218
+ page_content=' Automatic pulmonary lobe segmentation using deep learning.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
219
+ page_content=' arXiv Preprint arXiv:1903.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
220
+ page_content='09879.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
221
+ page_content=' Väänänen, S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
222
+ page_content=' P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
223
+ page_content=', Grassi, L.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
224
+ page_content=', Flivik, G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
225
+ page_content=', Jurvelin, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
226
+ page_content=' S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
227
+ page_content=', & Isaksson, H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
228
+ page_content=' (2015).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
229
+ page_content=' Generation of 3D shape, density, cortical thickness and finite element mesh of proximal femur from a DXA image.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
230
+ page_content=' Medical Image Analysis, 24(1), 125–134.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
231
+ page_content=' Williams et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
232
+ page_content=' (2023).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
233
+ page_content=' pyssam – a Python library for statistical modelling of biomedical shape and appearance.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
234
+ page_content=' Journal of Open Source Software, TBD, TBD.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
235
+ page_content=' https://doi.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
236
+ page_content='org/TBD 5' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/59E3T4oBgHgl3EQfQwmQ/content/2301.04416v1.pdf'}
8dFLT4oBgHgl3EQftC_m/content/tmp_files/2301.12150v1.pdf.txt ADDED
@@ -0,0 +1,540 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Wrapping pathways of anisotropic dumbbell
2
+ particles by giant unilamellar vesicles
3
+ Ali Azadbakht,†,∥ Billie Meadowcroft,‡,¶,∥ Thijs Varkevisser,†,§,∥ Anđela Šarić,‡ and
4
+ Daniela J. Kraft∗,†
5
+ †Soft Matter Physics, Huygens-Kamerlingh Onnes Laboratory, Leiden University, PO Box
6
+ 9504, 2300 RA Leiden, the Netherlands
7
+ ‡Institute of Science and Technology Austria, 3400 Klosterneuburg, Austria
8
+ ¶Department of Physics and Astronomy, Institute for the Physics of Living Systems,
9
+ University College London, London WC1E 6BT, United Kingdom
10
+ §Van der Waals-Zeeman Institute, Institute of Physics, University of Amsterdam, Science
11
+ Park 904, 1098 XH Amsterdam, Netherlands
12
+ ∥These authors contributed equally to this work.
13
+ E-mail: [email protected]
14
+ Abstract
15
+ Endocytosis is a key cellular process involved in the uptake of nutrients, pathogens
16
+ or the diagnosis and therapy of diseases. Most studies have focused on spherical ob-
17
+ jects, whereas biologically relevant shapes can be highly anisotropic. In this letter, we
18
+ use an experimental model system based on Giant Unilamellar Vesicles (GUVs) and
19
+ dumbbell-shaped colloidal particles to mimic and investigate the first stage of the pas-
20
+ sive endocytic process: engulfment of an anisotropic object by the membrane. Our
21
+ model has specific ligand-receptor interactions realized by mobile receptors on the vesi-
22
+ cles and immobile ligands on the particles. Through a series of experiments, theory
23
+ 1
24
+ arXiv:2301.12150v1 [cond-mat.soft] 28 Jan 2023
25
+
26
+ and molecular dynamics simulations, we quantify the wrapping process of anisotropic
27
+ dumbbells by GUVs and identify distinct stages of the wrapping pathway. We find that
28
+ the strong curvature variation in the neck of the dumbbell as well as membrane tension
29
+ are crucial in determining both the speed of wrapping and the final states.
30
+ The engulfment of objects through the cell membrane is critical for endocytic processes
31
+ such as phagocytosis1–3 and receptor-mediated endocytosis. The latter is often exploited by
32
+ viruses for cell entry and proliferation4 and key to nanomedical applications such as drug
33
+ delivery and imaging.5 To single out receptor-mediated effects from active mechanisms in-
34
+ volved in the engulfment,6 simplified passive model systems can be employed, which recently
35
+ led to a conclusive understanding of the wrapping of spherical objects.7,8 However, biological
36
+ objects such as bacteria and viruses4,9,10 as well as nanoparticles relevant for applications
37
+ in nanomedicine but also nanotoxicology11 often posses non-spherical shapes. Moreover, in
38
+ vitro experiments with nanoparticles and simulations have shown that the size and shape
39
+ influence their likelihood to be taken up by endocytosis.6,12–17
40
+ The wrapping pathways of spheres at sufficiently low membrane tensions have been shown
41
+ to be a continuous transition from attached to fully wrapped, occurring either spontaneously
42
+ or after activation.7,8,18 In contrast, anisotropic particles such as ellipsoids and rods, are
43
+ expected to reorient during the wrapping process or become trapped in metastable states
44
+ due to their varying curvature.19–27 The aspect ratio of these particles as well as the degree of
45
+ rounding of their tip were the key parameters affecting the wrapping orientation with respect
46
+ to the membrane and their metastable and stable states.24,27 Despite the extensive work in
47
+ theory and simulations and exciting observations on shape-dependence in phagocytosis,28 no
48
+ experimental work has investigated the passive wrapping process of anisotropic particles by
49
+ lipid membranes and tested these predictions yet.
50
+ In this letter, we employ an experimental model system based on Giant Unilamellar
51
+ Vesicles (GUVs) and colloidal dumbbell particles to investigate the wrapping of micrometre-
52
+ sized anisotropic objects by lipid membranes. Our model system is designed to have mobile
53
+ 2
54
+
55
+ ligands on the vesicles and immobile receptors on the particles mimicking receptor-mediated
56
+ endocytotic systems.18,29,30 We quantify the wrapping pathways of anisotropic dumbbells
57
+ by lipid membranes and test if their initial orientation affects the final states. Molecular
58
+ dynamics simulations of the same system corroborate our experimental data, allowing us to
59
+ inspect the dynamics of the process that was inaccessible to experiment. We find that the
60
+ strong curvature variation in the neck of the dumbbell as well as membrane tension and not
61
+ their initial orientation are crucial in both determining the speed of wrapping and the final
62
+ states.
63
+ We investigate the wrapping process of anisotropic objects by a lipid membrane using a
64
+ model system consisting of GUVs and colloidal particles, (see Fig. 1a). We chose the simplest
65
+ object that features anisotropy: a dumbbell shaped colloidal particle that consists of two
66
+ equal sized spheres.
67
+ The colloid dumbbells were obtained from aggregating polystyrene
68
+ spheres with diameter ds=0.98± 0.03 µm31 by briefly lowering the pH to 5.3 and then
69
+ quenching the process by increasing the pH to 8.6.32 This process yielded 5-10% dimers with
70
+ a long axis of 1.96 ± 0.06 µm and a short axis of 0.98 ± 0.03 µm. GUVs were prepared by
71
+ electroswelling from 97.5% w/w 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC).
72
+ To realize strong ligand-receptor mediated binding we doped the GUVs with 2% w/w 1,2-
73
+ dioleoyl-sn-glycero-3-phosphoethanolamine-N-[biotin-2000] (DOPE-PEG2000-Biotin) and the
74
+ dumbbells with 2.2×103/µm2 NeutrAvidin following,31 see Fig. 1b and c and see particle
75
+ functionalization and quantification of binding affinity in Supporting Information. We sup-
76
+ press electrostatic interactions by working in 50 mM Phosphate Buffered Saline, and achieve
77
+ colloidal stability by coating the dumbbells with polyethyleneglycol (PEG5000). Imaging of
78
+ the position and orientation of the dumbbells and membranes in three dimensions was made
79
+ possible by dying the colloids with BODIPY, represented by a green color throughout the
80
+ manuscript, as well as including 0.5% w/w 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine-
81
+ N-(lissamine rhodamine B sulfonyl) (DOPE-Rhodamine) into the GUVs, represented by a
82
+ magenta color. See Fig. 1c. Confocal stacks and image sequences were acquired with an
83
+ 3
84
+
85
+ inverted Nikon TI-e microscope, equipped with a 60x (NA 1.2) objective and A1-R scan
86
+ head. 2D image sequences were taken at 59 fps, which enables tracking of the dumbbells in
87
+ real time. Experimental details are described in the Supporting Information.
88
+ To initiate the wrapping process, we used optical tweezers to bring dumbbell particles in
89
+ contact with the GUV. They subsequently diffused on the GUV surface before suddenly and
90
+ quickly becoming wrapped, a process that took between a few seconds and a few minutes
91
+ depending on membrane tension, see Figure 1e and Movie S1. To capture the wrapping pro-
92
+ cess with high speed, we adjusted the focal height during acquisition of the image sequence.
93
+ After wrapping, the dumbbell continued to diffuse on the inside of the vesicle.
94
+ We quantify the wrapping process of a dumbbell by measuring the angle θ between the
95
+ major axis of the dumbbell and surface normal of the GUV and distance d of the dumbbell
96
+ with respect to the undistorted surface of the GUV, see Figure 1d. We inferred the 3D
97
+ position of the dumbbell from the position of its lobes with respect to the GUV. To improve
98
+ the accuracy of tracking, particles were tracked only when their center of mass was between
99
+ -0.8R<z<0.8R, and when both lobes were in focus. Details are described in the Supporting
100
+ Information.
101
+ We show confocal microscopy snapshots of a typical wrapping pathway in Figure 1 e,
102
+ and quantitative data of θ and d for exemplary pathways in Figure 2a and b. Surprisingly,
103
+ we find that the dumbbells end up in one of two states even though they start from different
104
+ initial orientations: (1) both lobes are either being fully wrapped (Fig. 2a), or (2) a single
105
+ lobe is being wrapped, such that the dumbbell is engulfed up to its waist by the membrane
106
+ (Fig. 2b). The green-blue points in Fig. 2a and b represent dumbbells attached almost
107
+ parallel to the membrane at the beginning of the process, whereas the yellow-red points
108
+ represent dumbbells attached roughly perpendicular with respect to the membrane initially.
109
+ Other starting orientations also lead to either a fully wrapped or a half wrapped dumbbell,
110
+ but the probability for reaching either state was influenced by the initial position as we will
111
+ discuss below.
112
+ 4
113
+
114
+ θ
115
+ d
116
+ I
117
+ II
118
+ III
119
+ IV
120
+ V
121
+ VI
122
+ VII
123
+ r
124
+ z
125
+ R
126
+ b)
127
+ d)
128
+ t=10 s
129
+ t=30 s
130
+ t=50 s
131
+ t=70 s
132
+ t=90 s
133
+ t=110 s
134
+ a)
135
+ c)
136
+ e)
137
+ {
138
+ y [µm]
139
+ Figure 1: Experimental setup to quantitatively measure the wrapping process of
140
+ a dumbbell colloid by a GUV a) 3D confocal reconstruction of a GUV in magenta and
141
+ a dumbbell particle in green with an indication of the relative height z from the equator
142
+ of the GUV, radius of GUV R, and cross section radius of the vesicle at the location of
143
+ the dumbbell, r. b) Detailed schematic of ligand-receptor based binding scheme between the
144
+ dumbbell and GUV. I-DOPC lipid II-DOPE lipid III -Biotin IV -NeutrAvidin V-Rhodamine,
145
+ VI -Polyethylene glycol (PEG) VII -Polystyrene particle. (Not to scale) c) Representative
146
+ confocal images reconstructed from two channels, (1) dumbbell excited by 488 nm laser
147
+ light and emission collected between 500-550 nm (2) GUV excited by 561 nm laser light
148
+ and emission collected in 580-630 nm (scale bar 1µm). d) Schematic representation of the
149
+ parameters d and θ used for the quantitative description of the wrapping process. e) Time
150
+ series of snapshots of confocal images of a dumbbell being wrapped by a vesicle (scale bar
151
+ 4µm).
152
+ If the dumbbell is oriented parallel to the membrane initially (θ ≈ 90◦ and proceeds to
153
+ a fully wrapped state, then it tilts in the first part of the engulfment process to about 60◦.
154
+ Subsequently, its CoM moves inward to almost d ≈ 1.5ds from the undisturbed membrane
155
+ contour, before returning to a more parallel orientation and an insertion depth about d ≈
156
+ 0.7ds. This overshooting and recoil is similar to that observed for spheres previously.8,33
157
+ If the dumbbell initially is roughly perpendicular the membrane, it first becomes oriented
158
+ more precisely perpendicular until it is covered halfway (d = 0 and θ ≈ 10◦) before being
159
+ wrapped further and finally ending in a more parallel orientation at a similar distance from
160
+ 5
161
+
162
+ 25
163
+ 20
164
+ wr]
165
+ 15
166
+ 10
167
+ 5
168
+ X [um]
169
+ 0
170
+ 0
171
+ 20
172
+ 5
173
+ 10
174
+ 15
175
+ 20
176
+ 25
177
+ 302 μm
178
+ 0:00:09.8112 μm
179
+ 0:00:29.9322 μm
180
+ 0:00:49.4992 μm
181
+ 0:01:09.9972 μm
182
+ 0:01:30.0882 μm
183
+ 0:01:49.789the undisturbed membrane as the initially parallel dumbbells.
184
+ For final states where one lobe is being wrapped only, an initially perpendicular dumbbell
185
+ first reorients more parallel before becoming engulfed until its waist while becoming perpen-
186
+ dicular again. An initially parallel dumbbell proceeds to reorient perpendicular while being
187
+ engulfed, see Fig. 2b. The gap in the yellow-red trace at θ ≈ 55◦ and d=0.5 µm was caused
188
+ by the dumbbell going through an orientation that was filtered out for accuracy as described
189
+ above.
190
+ To obtain more quantitative results for the dynamics of the system we carried out coarse-
191
+ grained (CG) molecular dynamics (MD) simulations of anisotropic dumbbell particles being
192
+ wrapped by a membrane. Besides the advantage of easily measuring dynamic properties,
193
+ in these simulations we are also able to control the size of the vesicle and dumbbell, the
194
+ membrane tension and the interaction strength between dumbbell and membrane and thus
195
+ probe a wider parameter space than is available to experiments.
196
+ The membrane is modelled using a one particle thick fluid surface developed by Yuan
197
+ et al34 which reproduces the mechanical properties associated with biological membranes.35
198
+ Using this model, we simulate spherical membrane vesicles and change the membrane tension
199
+ by the addition of small solute particles on the inside and outside of the vesicle.36 The solute
200
+ particles only interact via volume exclusion and produce a pressure force when the inside and
201
+ outside concentrations are different. The dumbbell colloid is then placed on the membrane in
202
+ either a vertical or horizontal initial condition and due to the attractive interaction between
203
+ the membrane beads and the dumbbell, the dumbbell is slowly wrapped and engulfed by the
204
+ vesicle. Details can be found in the Supporting Information.
205
+ The results obtained from simulations show qualitatively similar behavior as in the exper-
206
+ iments, see Figure 2. Again, both final states, i.e. i) one lobe attached and ii) fully engulfed,
207
+ could be reached from any initial position, and the pathway they took was influenced by the
208
+ initial orientation. Interestingly, our simulations suggest that the initial position strongly
209
+ influences the first part of the wrapping process and to a lesser degree the second half,
210
+ 6
211
+
212
+ which is observed to be similar for both extreme initial orientations. The observation that
213
+ the wrapping pathways from different initial positions can result in the same final position
214
+ shows that there is an energy minimum for the GUV-dumbbell system independent of the
215
+ initial position of the dumbbell. In all observed pathways towards the fully wrapped state,
216
+ the dumbbell particle tilts during the engulfment suggesting that this requires less bending
217
+ energy.
218
+ A similar reorientation upon wrapping was observed for linear aggregate of particles37 and
219
+ elongated ellipsoids.21,25–27 Ellipsoids have been found to become first adhered by the side,
220
+ before rotating to the tip upon being wrapped by the membrane.25 For sphero-cylindrical
221
+ particles that were initially touching with their tip, a rotation-mediated wrapping was also
222
+ seen,17,23 which can rotate the particle from a standing to a lying position at high aspect
223
+ ratios. The first point of contact has been predicted to be crucial for the ultimate fate of
224
+ a non-spherical particle.26,27 In contrast, for the dumbbell particles used here rotation is
225
+ not driven by a variation of particle curvature, but primarily by thermal fluctuations and
226
+ possibly inhomogeneities in the ligand coating density, because of the constant curvature
227
+ of the constituent spheres of the dumbbells. The only region of curvature variation is the
228
+ dumbbell neck, which we will show to play a crucial role in the wrapping.
229
+ From the many wrapping processes we observed in experiments and simulations, we iden-
230
+ tified a number of key intermediate states during the engulfment that ultimately determined
231
+ the final state. A decisive event during the wrapping of the first lobe is whether the second
232
+ lobe gets bound to the membrane. This is always the case if the particle starts out being
233
+ perfectly parallel and thus with both lobes attached (Figure 3A3). If the particle initially
234
+ is attached with a single lobe (3A1 and A2), however, tilting during the engulfment may
235
+ attach the second lobe (3B). In principle, since one lobe is spherical one may expect engulf-
236
+ ment to proceed uniformly, not inducing or requiring any tilt. However, any inhomogeneity
237
+ in the coating density of the ligands on the dumbbells, as well as thermal fluctuations will
238
+ tilt the particle and may induce contact of the second lobe to the membrane. Since biotin-
239
+ 7
240
+
241
+ a)
242
+ b)
243
+ c)
244
+ d)
245
+ Experiment
246
+ Simulations
247
+ Figure 2: Quantitative wrapping pathway of dumbbell particles by for GUVs.
248
+ Tilt angle θ and distance d of the dumbbell from the vesicle surface obtained from a,b)
249
+ experiments and c,d) simulations as a function of time. In all panels, green-blue pathways
250
+ indicate dumbbells starting from a vertical position with respect to the vesicle surface, and
251
+ yellow-red pathways indicate dumbbells that initially start almost horizontally with respect
252
+ to the membrane.
253
+ Time is indicated by color, specified by colorbars for each panel.
254
+ a)
255
+ Experimentally obtained pathways for a dumbbell initially oriented parallel or perpendicular
256
+ to the membrane surface to a fully wrapped end state. Each data point represents an average
257
+ over 1s. b) Experimentally obtained pathways taken by a dumbbell initially oriented parallel
258
+ or perpendicular to the membrane surface to the half-wrapped end state. Each data point
259
+ represents an average over 5s. c) Simulations of pathways for a dumbbell initially oriented
260
+ parallel and perpendicular to the membrane surface to the fully-wrapped end state. This
261
+ was the most common stable state with ∼ 90% of dumbbells reaching this end state. d)
262
+ Simulation of pathways for a dumbbell initially oriented parallel and perpendicular to the
263
+ membrane surface to the half-wrapped end state. a-d) Circle size indicates the number of
264
+ images used for the average. Simulation time is in expressed in ∆T = 0.01τ0, τ0 being MD
265
+ unit of time.
266
+ Neutravidin interactions are essentially irreversible at room temperature, attachment of the
267
+ second lobe always precludes achieving a final state where only one lobe is wrapped. If
268
+ the second lobe does not attach, the single-wrapped lobe state is reached (3D). Otherwise,
269
+ the dumbbell will wrap both lobes consecutively, either in a symmetric fashion (3E2) or in
270
+ an asymmetric way (3E1), leading to the fully wrapped state. The symmetric wrapping is
271
+ unstable, and eventually leads to Fig. 3F in which both lobes are covered. The angle the
272
+ dumbbell makes with the membrane after wrapping completed can vary. In this end state, a
273
+ small neck connected the fully wrapped dumbbell at one lobe with the vesicle, see Fig. 3F.
274
+ To quantify the time evolution, we measured the transition times between the different
275
+ 8
276
+
277
+ wrapping states. Membrane tension was found to be crucial for the overall wrapping time, see
278
+ below, and therefore simulations were used for quantitative measurements of the transition
279
+ times and experiments for qualitative comparison. While the initial wrapping of the first
280
+ lobe in the simulations is almost equally fast for the different initial states (see Fig. 3G and
281
+ H), the wrapping slowed down significantly when the membrane was crossing the waist (Fig.
282
+ 3G and H). This signifies an energy barrier stemming from the high bending energy required
283
+ to adapt to the strong variation in curvature of the particle surface. For dumbbells with both
284
+ lobes attached, we observed slowing down at the waist (Fig. 3G). For dumbbells attached
285
+ with a single lobe only, the wrapping process stopped for a longer time at the waist (Fig. 3H).
286
+ We observed the same qualitative behavior in experiments, both for tense and floppy GUVs,
287
+ indicating that the bending energy required to continue wrapping largely exceeded the energy
288
+ gained from adhesion. In experiments, in less than 10% of the cases, we observed dumbbells
289
+ wrapped with one lobe (3D) to suddenly transition to the fully engulfed state within about
290
+ 10 minutes, but never observed this within the timescales used in simulations in line with
291
+ ref.13 The high bending energy costs at the waist and the significantly faster wrapping for
292
+ tilted dumbbells observed in both simulations and experiments suggest that wrapping a
293
+ tilted dumbbell is less energetically costly than one that is oriented perpendicular to the
294
+ membrane.25 The strong trapping at the waist also causes single-lobe wrapped dumbbells to
295
+ attain their stable insertion depth d without overshooting and recoil.
296
+ The probability of following a specific pathway and reaching one of the two final states
297
+ as qualitatively observed in experiments, depended on two factors: the membrane tension of
298
+ the GUV and the dumbbell’s angle θ0 with respect to the membrane’s surface normal during
299
+ the initial wrapping. The higher the surface tension of the GUV, the more likely it was for
300
+ the dumbbell to end up in situation 3D. Large fluctuations of the vesicle’s surface enabled
301
+ the dumbbell to attach to the non-wrapped lobe. The larger the angle θ in situation 3A2,
302
+ and thus the closer to the membrane it started out at the more likely it was for the dumbbell
303
+ to end up in situation 3B and hence E1.
304
+ 9
305
+
306
+ F
307
+ E1
308
+ E2
309
+ C
310
+ D
311
+ A3
312
+ A1
313
+ B
314
+ A2
315
+ A3
316
+ C
317
+ A1
318
+ F
319
+ E1
320
+ B
321
+ G
322
+ H
323
+ Experiments
324
+ Simulations
325
+ Simulations
326
+ Simulations
327
+ Figure 3: Overview of the observed wrapping pathways. A1-F) Confocal images of
328
+ the possible orientation of a dumbbell (All scale bars denote 1µm). Arrows indicate the
329
+ directions of the possible wrapping pathways, and dashed arrows illustrate transitions that
330
+ were rarely observed. G) Measurements of the time between the states for the horizontal
331
+ dumbbell starting position, given in simulation timesteps. H) Measurements of the time
332
+ between the states for the vertical dumbbell starting position.
333
+ The overall time as well as the transition between different stages in the wrapping strongly
334
+ depended on the membrane tension - both the initial tension as well as the tension at later
335
+ times which will increase because of the wrapping, see Figure 4. We experimentally measured
336
+ the membrane tension from the fluctuation spectrum of the lipid vesicle following ref.38 and
337
+ plot the time taken to complete wrapping as a function of membrane tension in Figure
338
+ 4a,b. We observed an increase in overall wrapping time with increasing initial membrane
339
+ tension in experiments (e.g. Figure 4a,b) and simulations (Figure 4c). However, the range
340
+ of tensions we could replicate in experiments and simulations was quite limited. To be able
341
+ to fully explore this effect, we extended a previously developed analytical theory describing
342
+ the time to wrap colloids,39,40 which was recently experimentally confirmed,8 and adapted
343
+ it to the shape of a dumbbell (Details of the theory can be found in the SI). In doing so we
344
+ 10
345
+
346
+ could explore the effect of tension on time to wrap the dumbbell for a range of theoretical
347
+ parameters. All the parameters used in the theory were taken directly from the experiment,
348
+ apart from the binding energy per area (W) and the microviscosity of the membrane (ηeff)
349
+ which are both discussed below.
350
+ For a given adhesion energy, we find that the time taken to fully wrap the dumbbell in-
351
+ creases non-linearly with the tension. With increasing adhesion energy, the wrapping process
352
+ becomes faster at the same tension, see Figure 4a. The adhesion energies in experiments
353
+ vary due to the distribution of binding sites between dumbbells18,31 which is also reflected
354
+ in that the experimental data points fall within a range of adhesion energies identified by
355
+ the theory. We note that only a small percentage of the NeutrAvidin sites that have been
356
+ added during synthesis contribute to the effective adhesion energy, as was found previously
357
+ in ref.18 Although fixed in the experiments, varying membrane microviscosity in the theory
358
+ also changes the time taken to wrap. Membrane microviscosity is a measure of how easily
359
+ the lipids slide past each other during rearrangement, and a higher microviscosity is linked
360
+ to a higher frictional force during colloid-membrane wrapping. The comparison between
361
+ the theoretical and experimental results allows us to estimate the membrane microviscosity,
362
+ which is experimentally inaccessible. We find that our experimental measurements best fit
363
+ the theoretical curves for a membrane microviscosity of ηeff ≈ 0.8 Pa·s, Figure 4b, about 10
364
+ times larger than the lower bound estimated in.8 However, the theory in ref8 consistently
365
+ over-estimated the wrapping speed as compared with experiments on spheres, so it could be
366
+ that the experiment microviscosity was larger than their theoretically predicted value.
367
+ Here we have developed the first model system to quantitatively study ligand-receptor
368
+ mediated endocytosis of an anisotropic object by making use of GUVs and colloidal dumbbell
369
+ particles. We followed and quantified their orientation θ and distance d with respect to the
370
+ membrane during wrapping using experiments and molecular dynamics simulations.
371
+ We
372
+ found that there are two final states: 1) only one lobe or 2) both lobes of the dumbbell are
373
+ fully wrapped by the membrane. The two states can be reached from any initial position
374
+ 11
375
+
376
+ a)
377
+ b)
378
+ c)
379
+ Figure 4: Measurement of the time required to fully wrap a dumbbell-shaped
380
+ particle as a function of membrane tension (a) Experimental data (points) and the-
381
+ oretical predictions (lines) for different membrane viscosity in the range of 0.4-1.6 Pa·s at
382
+ a fixed adhesion energy per unit of area of 0.76 µJ/m2. (b) Experimental data (points)
383
+ and theoretical predictions (lines) for different adhesion energy per unit area int he range
384
+ of 0.69-0.79 µJ/m2 at a fixed membrane viscosity of 0.8 Pa·s. c) Time to fully wrap the
385
+ dumbbell-shaped particle in simulations for a range of tensions <10 nN/m.
386
+ except when both lobes were attached initially which necessarily leads to full wrapping of
387
+ both lobes. However, the initial position influenced the pathway towards the final state. We
388
+ identified a number of key intermediate states during the wrapping that determine the final
389
+ state. Wrapping of one lobe was only found for high membrane tensions and if the other lobe
390
+ did not touch the membrane during engulfment. Using molecular dynamics simulations we
391
+ quantified the time required between key intermediate steps, with the slowest step being the
392
+ crossing of the highly curved neck region of the dumbbell. With simulations we confirmed
393
+ the experimentally-observed trend of time to wrap increasing for increasing tension, and
394
+ using analytical theory we estimated the membrane microviscosity.
395
+ Our results contribute to a better understanding of how shape affects endocytosis, nutri-
396
+ tion uptake, and bacterial evasion. Our choice of a simple anisotropic object, a dumbbell,
397
+ enabled a key insight: highly negatively curved regions may dominate the wrapping and
398
+ possibly even prevent full engulfment unless active processes are present. This suggests that
399
+ objects, such as certain viruses such as pox virus4that rely on endocytosis, may profit from
400
+ having a convex shape. Incorporation of active processes, such as those driven by actin or
401
+ ESCRT-III polymers, could provide further insights into how the competition between the
402
+ 12
403
+
404
+ passive and active processes affects wrapping.
405
+ Supporting Information
406
+ • "Supporting Information: Details of the experiments and simulations; experimental
407
+ materials used; methods employed for membrane preparation, particle functionaliza-
408
+ tion, experimental imaging, and quantification of ligands on dumbbells; details of quan-
409
+ tification of dumbbell wrapping and filters applied for the analysis; detail of theory used
410
+ for time taken to wrap a dumbbell"
411
+ • Movie S1: An example of an experimental video of the wrapping process of a dumbbell
412
+ colloid attached to a GUV, recorded with a confocal microscope at 59 frames per second
413
+ and 3× accelerated.
414
+ Movie S1: An example of an experimental video of the wrapping process of a dumbbell
415
+ colloid attached to a GUV, recorded with a confocal microscope at 59 frames per second and
416
+ 3× accelerated.
417
+ Acknowledgments
418
+ We sincerely thank Casper van der Wel for providing open-source packages for tracking, as
419
+ well as Yogesh Shelke for his assistance with PAA coverslip preparation and Rachel Doherty
420
+ for her assistance with particle functionalization. We are grateful to Felix Frey for useful
421
+ discussions on the theory of membrane wrapping. B.M. and A.Š. acknowledge funding by the
422
+ European Union’s Horizon 2020 research and innovation programme (ERC Starting Grant
423
+ No. 802960).
424
+ 13
425
+
426
+ References
427
+ (1) Douglas, T.; Young, M. Viruses: Making friends with old foes. Science 2006, 312,
428
+ 873–875.
429
+ (2) Manchester, M.; Singh, P. Virus-based nanoparticles (VNPs): Platform technologies
430
+ for diagnostic imaging. Advanced Drug Delivery Reviews 2006, 58, 1505–1522.
431
+ (3) Lewis, J. D.; Destito, G.; Zijlstra, A.; Gonzalez, M. J.; Quigley, J. P.; Manchester, M.;
432
+ Stuhlmann, H. Viral nanoparticles as tools for intravital vascular imaging. Nature
433
+ Medicine 2006, 12, 354–360.
434
+ (4) Schmidt, F. I.; Bleck, C. K. E.; Mercer, J. Poxvirus host cell entry. Current opinion in
435
+ virology 2012, 2, 20–27.
436
+ (5) Mitchell, M. J.; Billingsley, M. M.; Haley, R. M.; Wechsler, M. E.; Peppas, N. A.;
437
+ Langer, R. Engineering precision nanoparticles for drug delivery. Nature Reviews Drug
438
+ Discovery 2021, 20, 101–124, Number: 2 Publisher: Nature Publishing Group.
439
+ (6) Richards, D. M.; Endres, R. G. The mechanism of phagocytosis: Two stages of engulf-
440
+ ment. Biophysical Journal 2014, 107, 1542–1553.
441
+ (7) Spanke, H. T.; Style, R. W.; François-Martin, C.; Feofilova, M.; Eisentraut, M.;
442
+ Kress, H.; Agudo-Canalejo, J.; Dufresne, E. R. Wrapping of Microparticles by Floppy
443
+ Lipid Vesicles. Physical Review Letters 2020, 125, 1–9.
444
+ (8) Spanke, H. T.; Agudo-Canalejo, J.; Tran, D.; Style, R. W.; Dufresne, E. R. Dynamics
445
+ of spontaneous wrapping of microparticles by floppy lipid membranes. Physical Review
446
+ Research 2022, 4, 23080.
447
+ (9) Young, K. D. The selective value of bacterial shape. Microbiology and molecular biology
448
+ reviews 2006, 70, 660–703.
449
+ 14
450
+
451
+ (10) Levy, J. A.; Fraenkel-Conrat, H.; Owens, O. S. Virology, 3rd ed.; Benjamin Cummings:
452
+ Englewood Cliffs, N.J, 1994.
453
+ (11) Cho, K.; Wang, X.; Nie, S.; Chen, Z. G.; Shin, D. M. Therapeutic Nanoparticles for
454
+ Drug Delivery in Cancer. Clinical Cancer Research 2008, 14, 1310–1316.
455
+ (12) Aoyama, Y.; Kanamori, T.; Nakai, T.; Sasaki, T.; Horiuchi, S.; Sando, S.; Niidome, T.
456
+ Artificial viruses and their application to gene delivery. Size-controlled gene coating
457
+ with glycocluster nanoparticles. Journal of the American Chemical Society 2003, 125,
458
+ 3455–3457.
459
+ (13) Richards, D. M.; Endres, R. G. Target shape dependence in a simple model of receptor-
460
+ mediated endocytosis and phagocytosis. Proceedings of the National Academy of Sci-
461
+ ences of the United States of America 2016, 113, 6113–6118.
462
+ (14) Chithrani, B. D.; Ghazani, A. A.; Chan, W. C. Determining the size and shape de-
463
+ pendence of gold nanoparticle uptake into mammalian cells. Nano Letters 2006, 6,
464
+ 662–668.
465
+ (15) Chithrani, B. D.; Chan, W. C. Elucidating the mechanism of cellular uptake and re-
466
+ moval of protein-coated gold nanoparticles of different sizes and shapes. Nano Letters
467
+ 2007, 7, 1542–1550.
468
+ (16) Li, X. Size and shape effects on receptor-mediated endocytosis of nanoparticles. Journal
469
+ of Applied Physics 2012, 111, 2010–2014.
470
+ (17) Dasgupta, R.; Dimova, R. Inward and outward membrane tubes pulled from giant
471
+ vesicles. Journal of Physics D: Applied Physics 2014, 47, 282001.
472
+ (18) Van Der Wel, C.; Vahid, A.; ŠariĆ, A.; Idema, T.; Heinrich, D.; Kraft, D. J. Lipid
473
+ membrane-mediated attraction between curvature inducing objects. Scientific Reports
474
+ 2016, 6, 1–10.
475
+ 15
476
+
477
+ (19) Decuzzi, P.; Ferrari, M. The Receptor-Mediated Endocytosis of Nonspherical Particles.
478
+ Biophysical Journal 2008, 94, 3790–3797.
479
+ (20) Vácha, R.; Martinez-Veracoechea, F. J.; Frenkel, D. Receptor-mediated endocytosis of
480
+ nanoparticles of various shapes. Nano Letters 2011, 11, 5391–5395.
481
+ (21) Bahrami, A. H. Orientational changes and impaired internalization of ellipsoidal
482
+ nanoparticles by vesicle membranes. Soft Matter 2013, 9, 8642–8646.
483
+ (22) Yang, K.; Yuan, B.; Ma, Y.-q. Influence of geometric nanoparticle rotation on cellular
484
+ internalization process. Nanoscale 2013, 5, 7998–8006, Publisher: The Royal Society
485
+ of Chemistry.
486
+ (23) Huang, C.; Zhang, Y.; Yuan, H.; Gao, H.; Zhang, S. Role of nanoparticle geometry in
487
+ endocytosis: Laying down to stand up. Nano Letters 2013, 13, 4546–4550.
488
+ (24) Dasgupta, S.; Auth, T.; Gompper, G. Shape and Orientation Matter for the Cellular
489
+ Uptake of Nonspherical Particles. Nano Letters 2014, 14, 687–693, Publisher: Ameri-
490
+ can Chemical Society.
491
+ (25) Bahrami, A. H.; Raatz, M.; Agudo-Canalejo, J.; Michel, R.; Curtis, E. M.; Hall, C. K.;
492
+ Gradzielski, M.; Lipowsky, R.; Weikl, T. R. Wrapping of nanoparticles by membranes.
493
+ Advances in Colloid and Interface Science 2014, 208, 214–224.
494
+ (26) Tang, H.; Zhang, H.; Ye, H.; Zheng, Y. Receptor-Mediated Endocytosis of Nanoparti-
495
+ cles: Roles of Shapes, Orientations, and Rotations of Nanoparticles. Journal of Physical
496
+ Chemistry B 2018, 122, 171–180.
497
+ (27) Agudo-Canalejo, J. Engulfment of ellipsoidal nanoparticles by membranes: full descrip-
498
+ tion of orientational changes. Journal of Physics: Condensed Matter 2020, 32, 294001.
499
+ (28) Champion, J. A.; Mitragotri, S. Role of target geometry in phagocytosis. Proceedings
500
+ 16
501
+
502
+ of the National Academy of Sciences of the United States of America 2006, 103, 4930–
503
+ 4934.
504
+ (29) van der Wel, C.; Heinrich, D.; Kraft, D. J. Microparticle Assembly Pathways on Lipid
505
+ Membranes. Biophysical Journal 2017, 113, 1037–1046.
506
+ (30) Sarfati, R.; Dufresne, E. R. Long-range attraction of particles adhered to lipid vesicles.
507
+ Physical Review E 2016, 94, 2–7.
508
+ (31) Van Der Wel, C.; Bossert, N.; Mank, Q. J.; Winter, M. G.; Heinrich, D.; Kraft, D. J.
509
+ Surfactant-free Colloidal Particles with Specific Binding Affinity. Langmuir 2017, 33,
510
+ 9803–9810.
511
+ (32) Meester, V.; Verweij, R. W.; Van Der Wel, C.; Kraft, D. J. Colloidal Recycling: Re-
512
+ configuration of Random Aggregates into Patchy Particles. ACS Nano 2016, 10, 4322–
513
+ 4329.
514
+ (33) Dietrich, C.; Angelova, M.; Pouligny, B. Adhesion of Latex spheres to giant phospholipid
515
+ vesicles: Statics and dynamics. Journal de physique. II 1997, 7, 1651–1682.
516
+ (34) Yuan, H.; Huang, C.; Li, J.; Lykotrafitis, G.; Zhang, S. One-particle-thick, solvent-
517
+ free, coarse-grained model for biological and biomimetic fluid membranes. Phys. Rev.
518
+ E 2010, 82, 011905.
519
+ (35) Curk, T.; Wirnsberger, P.; Dobnikar, J.; Frenkel, D.; Šarić, A. Controlling cargo traf-
520
+ ficking in multicomponent membranes. Nano letters 2018, 18, 5350–5356.
521
+ (36) Vanhille-Campos, C.; Šarić, A. Modelling the dynamics of vesicle reshaping and scission
522
+ under osmotic shocks. Soft Matter 2021, 17, 3798–3806.
523
+ (37) Šarić, A.; Cacciuto, A. Mechanism of membrane tube formation induced by adhesive
524
+ nanocomponents. Physical review letters 2012, 109, 188101.
525
+ 17
526
+
527
+ (38) Pécréaux, J.; Döbereiner, H. G.; Prost, J.; Joanny, J. F.; Bassereau, P. Refined contour
528
+ analysis of giant unilamellar vesicles. European Physical Journal E 2004, 13, 277–290.
529
+ (39) Agudo-Canalejo, J.; Lipowsky, R. Critical particle sizes for the engulfment of nanoparti-
530
+ cles by membranes and vesicles with bilayer asymmetry. ACS Nano 2015, 9, 3704–3720.
531
+ (40) Frey, F.; Ziebert, F.; Schwarz, U. S. Stochastic dynamics of nanoparticle and virus
532
+ uptake. Physical Review Letters 2019, 122, 88102.
533
+ 18
534
+
535
+ Graphical TOC Entry
536
+ time
537
+ Experiment
538
+ Simulation
539
+ 19
540
+
8dFLT4oBgHgl3EQftC_m/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
9dFLT4oBgHgl3EQfuS8F/content/tmp_files/2301.12154v1.pdf.txt ADDED
@@ -0,0 +1,1585 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+ 1
3
+ Critical resolved shear stresses for slip and twinning in Mg-Y-Ca alloys and their
4
+ effect on the ductility
5
+ Mingdi Yua, Yuchi Cuib, Jingya Wanga,*, Yiwen Chena, Zhigang Dingc, Tao Yinga,
6
+ Javier Llorcad,e,*, Xiaoqin Zenga,*
7
+ a National Engineering Research Center of Light Alloy Net Forming and State Key
8
+ Laboratory of Metal Matrix Composites, Shanghai Jiao Tong University, Shanghai
9
+ 200240, PR China.
10
+
11
+ b School of Materials Science and Engineering, Shanghai Jiao Tong University,
12
+ Shanghai, 200240, PR China.
13
+
14
+ c Nano and Heterogeneous Materials Center, School of Materials Science and
15
+ Engineering, Nanjing University of Science and Technology, Nanjing, Jiangsu
16
+ 210094, China.
17
+
18
+ d IMDEA Materials Institute, 28906 Getafe, Madrid, Spain.
19
+
20
+ e Department of Materials Science, Polytechnic University of Madrid, E. T. S. de
21
+ Ingenieros de Caminos, 28040 Madrid, Spain.
22
+
23
+ * Corresponding author. E-mail address: [email protected].
24
+ * Corresponding author. E-mail address: [email protected].
25
+ * Corresponding author. E-mail address: [email protected].
26
+
27
+ Abstract:
28
+ The deformation mechanisms of an extruded Mg-5Y-0.08Ca (wt. %) alloy were
29
+ analyzed by means of micropillar compression tests on single crystals along different
30
+ orientations -selected to activate specific deformation modes- as well as slip trace
31
+ analysis, transmission electron microscopy and transmission Kikuchi diffraction. The
32
+ polycrystalline alloy presented a remarkable ductility in tension (~32%) and negligible
33
+
34
+
35
+ 2
36
+ differences in the yield strength between tension and compression. It was found that the
37
+ presence of Y and Ca in solid solution led to a huge increase in the CRSS for <a> basal
38
+ slip (29 ± 5 MPa), <c+a> pyramidal slip (203 ± 7 MPa) and tensile twin nucleation
39
+ (above 148 MPa), while the CRSS for <a> prismatic slip only increases up to 105 ± 4
40
+ MPa. The changes in the CRSS for slip and tensile twinning in Mg-Y-Ca alloys
41
+ expectedly modify the dominant deformation mechanisms in polycrystals. In particular,
42
+ tensile twinning is replaced by <a> prismatic slip during compressive deformation
43
+ along the a-axis. The reduction of twinning (which generally induces strong anisotropy
44
+ in the plastic deformation in textured alloys), and the activation of <a> prismatic slip
45
+ (which provides an additional plastic deformation mechanism with limited hardening)
46
+ were responsible for the large tensile ductility of the alloy.
47
+
48
+ Keywords: Mg-Y-Ca alloys; micropillar compression; critical resolved shear stress;
49
+ plastic anisotropy; tension-compression asymmetry; tensile ductility.
50
+
51
+ 1. Introduction
52
+ Pure Mg and Mg alloys generally present poor ductility and formability, especially
53
+ at room temperature (Huang et al., 2022; Sun et al., 2019; Tang et al., 2022; Yaghoobi
54
+ et al., 2022). As a result, forming of rolled sheets and extruded bars becomes difficult
55
+ and limits the application of wrought Mg alloys in different industrial sectors (Li and
56
+ Fang, 2022). Thus, understanding the origin of the lack of ductility and formability is
57
+ of paramount importance to develop new Mg alloys that overcome these limitations.
58
+ The poor ductility of Mg alloys is primarily traced to its low-symmetry hexagonal
59
+ closed packed (HCP) lattice structure, which results in very large differences in the
60
+ critical resolved shear stress (CRSS) between basal and non-basal slip systems as well
61
+ as in the easy activation of tensile twinning (Lee et al., 2018). Plastic deformation in
62
+ pure Mg is initially accommodated by <a> basal slip, which only provides two
63
+ independent slip systems (Partridge, 1967). This process leads to the development of a
64
+ strong basal texture during rolling and extrusion. Moreover, plastic deformation along
65
+ the c-axis (which is necessary to activate five independent slip systems to fulfil the von-
66
+
67
+
68
+ 3
69
+ Mises criterion for homogeneous plastic deformation) is absorbed by tensile twinning,
70
+ which is triggered at much lower CRSS than that necessary to produce <c+a> pyramidal
71
+ slip (Graff et al., 2007; Sukedai and Yokoyama, 2010). However, the plastic strain
72
+ associated with tensile twinning is very limited (at most 7%), moreover, tensile twining
73
+ is a polar mechanism that only occurs when the stress along the c-axis of the crystal is
74
+ tensile (Mayama et al., 2011). This leads to a large buildup of stresses to activate <c+a>
75
+ pyramidal slip in grains that are not suitably oriented for twinning and/or that cannot
76
+ accommodate more plastic deformation by twinning (Obara et al., 1973; Reed-Hill and
77
+ Robertson, 1957). The stress concentrations in these grains facilitate the nucleation of
78
+ cracks and limit the ductility (Zhang et al., 2022). Moreover, huge differences in the
79
+ flow stress and the strain hardening rate between tension and compression appear in
80
+ textured microstructures, which also lead to fracture during bending and forming
81
+ operations (Agnew and Duygulu, 2005; Basu et al., 2021).
82
+ The strategies to improve ductility and formability of Mg alloys have been directed
83
+ towards promoting the activation of multiple slip, including non-basal <a> and non-
84
+ basal <c+a> slip, and to suppress deformation twinning. Multiple slip leads to more
85
+ homogeneous plastic deformation and limits texture development during rolling and
86
+ extrusion while twinning promotes plastic anisotropy in textured microstructures
87
+ (Ahmad et al., 2019; G. Liu et al., 2017; Zhang et al., 2016a). For instance, precipitation
88
+ hardening in Mg-Zn alloys leads to large enhancements in the CRSS for basal (Alizadeh
89
+ and LLorca, 2020; Chun and Byrne, 1969; Wang and Stanford, 2015) and pyramidal
90
+ slip (Alizadeh et al., 2021) and, thus, to an important reduction in the pyramidal-to-
91
+ basal CRSS ratio. Nevertheless, the large increase in flow stress inherently decreases
92
+ the ductility due to the strong accumulation of geometrically necessary dislocations
93
+ around the precipitates (Rosalie et al., 2012). In addition, precipitates also increase the
94
+ CRSS for twin growth but do not affect the CRSS for twin nucleation (Wang et al.,
95
+ 2019b). As the latter is normally higher than the former, the presence of precipitates do
96
+ not contribute to hinder the development of twinning. The only difference induced by
97
+ the precipitates is a larger number of smaller twins, as compared to the precipitate-free
98
+ condition (Stanford et al., 2012). Thus, precipitate is not very efficient to enhance the
99
+
100
+
101
+ 4
102
+ ductility of Mg alloys (Fu et al., 2019; Jain et al., 2010).
103
+ Strategies based on solid solution hardening have been more successful to improve
104
+ the ductility of Mg alloys if the alloying elements are properly chosen. For instance,
105
+ Sandlöbes et al. (Sandlöbes et al., 2013, 2012, 2011) reported that the addition of 3 wt. %
106
+ Y led to Mg alloys with a tensile ductility > 25 %, which was associated with the
107
+ presence of a large density of <c+a> pyramidal dislocations in the deformed sample.
108
+ This behavior was mainly attributed to a reduction in the ratio between the CRSS of the
109
+ < c+a > pyramidal slip and < a > basal slip, which was ~3.2 according to in situ high
110
+ energy X-ray diffraction tests (Huang et al., 2018; Wang et al., 2018) and ~2.8-4.8 from
111
+ micropillar compression tests (Wu et al., 2020). Large ductility and formability are not
112
+ achieved, however, by the addition of other elements in solid solution (such as Al or Zn)
113
+ because the pyramidal-to-basal CRSS ratio in these alloys are > 10 (Li et al., 2021a;
114
+ Wang et al., 2020). Zhu et al., (2019) found that the addition of 0.47 wt. % of Ca in
115
+ solid solution enhanced the activity of <a> prismatic and pyramidal I dislocations as
116
+ well as the cross-slip between basal and non-basal slip planes, improving the tensile
117
+ ductility to ~18 % in a Mg-0.47 Ca (wt. %) alloy. And several authors reported a large
118
+ improvement in the ductility of binary Mg-Zn and Mg-Al alloys through the addition
119
+ of small amount of Ca (Hofstetter et al., 2015; Sandlöbes et al., 2017; Wang et al.,
120
+ 2021b). This behavior was supported by our recent micropillar compression tests that
121
+ showed that the addition of Ca to Mg-Zn alloys reduced the pyramidal-to-basal CRSS
122
+ ratio values, that were similar to those found in Mg-Y alloys (Wang et al., 2021a).
123
+ Finally, Wu et al., (2018) showed that the presence of Y and Ca reduces the energy for
124
+ cross-slip/double cross-slip of <c+a> pyramidal dislocations, leading to new dislocation
125
+ loops which accommodate plastic deformation. In contrast, the cross-slip is inhibited in
126
+ pure Mg (or in Mg-Al and Mg-Zn alloys) (Wu et al., 2018), by the favorable
127
+ dissociation of edge pyramidal <c+a> dislocation segments into sessile segments in the
128
+ basal plane.
129
+ Regarding the effect of solid solution on tensile twinning, several investigations
130
+ reported an increase in the CRSS for twin nucleation and growth with the addition of
131
+ Al (Wang et al., 2020), Zn (Li et al., 2021a), Y (Li et al., 2021b) as well as Ca to Mg-
132
+
133
+
134
+ 5
135
+ Zn alloys (Wang et al., 2021a). However, the CRSS for twin nucleation and growth
136
+ were lower than that for <c+a> pyramidal slip in the corresponding alloy, thus, tensile
137
+ twinning was still preferred over pyramidal slip to accommodate plastic deformation in
138
+ grains suitable oriented for twinning. In addition, the addition of 4Y (wt. %) could
139
+ significantly suppress the tensile twinning (with CRSS larger than 113 MPa) and
140
+ promote the <c+a> dislocations (with CRSS around 106 MPa) (Wu et al., 2020). The
141
+ results summarized above point to the beneficial effects of Y and Ca in solid solution
142
+ to reduce the plastic anisotropy of Mg. Thus, the co-addition of Ca and Y is expected
143
+ to promote the homogeneous deformation and improve the plastic deformability of Mg
144
+ alloys, taking advantages of the significant suppression effect of Y on the tensile
145
+ twinning, the promotion effect of Ca on the non-basal <a> slips, simultaneously the
146
+ positive effect of Ca and Y on the activation of the <c+a> slips. Ca enhances the
147
+ activation of <a> prismatic and <a> pyramidal slip while Y has similar effects on <c+a>
148
+ slip. Moreover, experimental results on the tensile behavior of an extruded Mg – 2.4
149
+ wt. % Y – 0.3 wt. % Ca (Zhou et al., 2013) showed a very large tensile elongation
150
+ (~37 %) but there is not information available in the literature -to the authors’
151
+ knowledge- on the concurrent effects of Y and Ca in solid solution on the dominant
152
+ deformation mechanisms and this is the main objective of this investigation. Thus, the
153
+ CRSS for different slip systems and twinning was determined in a Mg-Y-Ca alloy from
154
+ micropillar compression tests in single crystals with different orientations. The
155
+ deformation mechanisms were ascertained from slip trace analysis in the scanning
156
+ electron microscope (SEM), transmission electron microscopy (TEM) observations of
157
+ the dislocations as well as transmission Kikuchi diffraction (TKD). This information
158
+ was used to rationalize the excellent ductility of Mg-Y-Ca and to provide guidelines to
159
+ design novel Mg alloys with improved ductility and formability.
160
+
161
+ 2. Materials and experimental techniques
162
+ 2.1 Materials
163
+ The Mg-Y-Ca alloy was prepared from pure Mg (99.99 wt. %), Mg-30 Ca (wt. %)
164
+ and Mg-30 Y (wt. %) master alloys in a resistance furnace under a protective
165
+
166
+
167
+ 6
168
+ atmosphere of CO2 and SF6. The actual chemical composition of the ingot, obtained by
169
+ inductively coupled plasma atomic emission spectroscopy, was Mg-5Y-0.08Ca (wt. %).
170
+ The cast alloy was solution treated at 400 ℃ for 12 h, followed by extrusion at 300 ℃
171
+ with an extrusion ratio of ~ 18:1. Afterwards, parallelepipedal samples of 10×10×5 mm3
172
+ were cut from the extruded specimens and homogenized at 550 ℃ for 20 days within
173
+ quartz capsules filled with Ar to induce grain growth.
174
+ 2.2 Experimental techniques
175
+ Tensile and compressive tests were carried out along the extrusion direction in
176
+ polycrystalline specimens at crosshead speed of 0.5 mm/min, using a universal testing
177
+ machine (Z100-TEW) at room temperature. The dimensions of the gage section of the
178
+ dog-bone tensile specimens were 18×3.4×1.4 mm3 (length × width × thickness), while
179
+ cylindrical specimens of 8 mm in diameter and 12 mm in length were used in the
180
+ compression tests. Deformation was measured with an extensometer and 3 specimens
181
+ were tested in each condition.
182
+ The crystallographic orientation of the grains in the sample was characterized via
183
+ electron back-scattered diffraction (EBSD) in a Tescan Mira-3 SEM with an Oxford
184
+ Instruments Nordlys EBSD detector at an accelerating voltage of 20 kV. The surface of
185
+ the sample was mechanically ground using abrasive SiC papers with a grit size of 1200,
186
+ 2000, 3000, 5000 and 7000. Subsequently, the sample surface was electropolished in
187
+ an ethanol solution with 10 (vol. %) perchloric acid at -30 ℃ and 30 V for 90 s to
188
+ remove the surface damage induced by grinding and reveal the grain boundaries. The
189
+ EBSD data were analyzed using the Channel 5 software and the Oxford Instruments
190
+ AZtec Nanoanalysis software package v6.0 along with AZtec Crystal. Several grains
191
+ whose orientations were appropriate to active different deformation modes were
192
+ selected to mill the micropillars.
193
+ Micropillars of 5 × 5 μm2 square cross and an aspect ratio 2:1 were milled from
194
+ the selected grains using a FEI Helios G4 UX Focused Ion Beam (FIB)/SEM dual beam
195
+ microscope operated at 30 kV. These dimensions are known to minimize size effects
196
+ during mechanical deformation while the time and effort to mill each micropillar are
197
+ reasonable (Wang et al., 2021a). An initial ion current of 9.3 nA was used to remove the
198
+
199
+
200
+ 7
201
+ surrounding material and it was reduced to 2.5 nA when the beam was getting closer to
202
+ the actual dimensions of the micropillar. A final ion current of 80 pA was used in the
203
+ final polishing step to minimize the surface damage due to Ga+ ion-implantation. The
204
+ final taper of the micropillars was < 1.5°.
205
+ Micropillar compression tests were performed in ex situ using a Hysitron
206
+ Triboindenter TI950 system though a diamond flat punch of 10 μm in diameter. All the
207
+ tests were conducted under displacement control up to a maximum strain of 10 % at a
208
+ nominal strain rate of 10-3 s-1. The experimental displacement was corrected to account
209
+ for the elastic deflection of the matrix material beneath the micropillars following the
210
+ Sneddon correction (Sneddon, 1965). To this end, the elastic modulus of each grain was
211
+ determined via the nanoindentation method with a Berkovich tip in the same grain
212
+ where the micropillar was milled. More details about micropillar manufacturing and the
213
+ compression set-up can be found in (Sneddon, 1965; Wang et al., 2021a).
214
+ The engineering stress-strain curves were obtained from the load and the corrected
215
+ elastic deflection of the micropillar using the initial cross-sectional area and the height
216
+ of the micropillars measured in the SEM. The yield stress, σy , was determined from
217
+ the loss of linearity in the stress-strain curve following the methodology described in
218
+ (Alizadeh and LLorca, 2020; Maaß et al., 2009). From this information, the CRSS of
219
+ the active slip system was determined as
220
+ CRSS = SF × σy (1)
221
+ where SF is the Schmid factor of the corresponding slip system, computed from the
222
+ crystallographic orientation of each crystal (Table 1).
223
+ The slip traces on the top and lateral surfaces of the deformed micropillars were
224
+ characterized in a Tescan Mira-3 SEM to ascertain the active slip planes. The active slip
225
+ plane and direction were identified from the micropillar orientation using VESTA
226
+ software (Momma and Izumi, 2008). Moreover, TEM and TKD were used to determine
227
+ the dislocation activity and the orientation of the micropillar after deformation. To this
228
+ end, a thin lamella was lifted-out along the loading direction from the deformed pillars
229
+ and thinned to < 100 nm in thickness using FIB. The TKD maps were collected in a
230
+
231
+
232
+ 8
233
+ Tescan Mira-3 SEM at 30 kV with a step size of 20 nm. The TEM observations were
234
+ carried out using a Talos F200X G2 microscope operated at 200 kV. The two-beam
235
+ condition was applied to obtain dislocation contrast. Moreover, the “g·b” visibility
236
+ criterion was used to identify the types of dislocation, i.e., the dislocation is in contrast
237
+ when g!⃗ · b!⃗ ≠ 0, where g!⃗ is the diffraction vector and b!⃗ the Burgers vector.
238
+ 2.3 First-principles calculations
239
+ In order to study the influence of Y and/or Ca atoms on the deformation
240
+ mechanisms in Mg alloys, the generalized stacking fault energy (GSFE) curves of
241
+ different slip systems were calculated via the first-principles calculations using the
242
+ Vienna Ab initio Simulation Package (VASP) (Kresse and Furthmüller, 1996). The
243
+ exchange-correlation function was described using the generalized gradient
244
+ approximation (GGA) with the Perdew-Burke-Ernzerholf functional (PBE), based on
245
+ the projector augmented wave (PAW) (Blöchl, 1994) method.
246
+ A supercell with 12-layers containing 48 atoms was defined for different slip
247
+ systems, as indicated in Fig. 1. Each supercell was separated by 15 Å vacuum to
248
+ eliminate the influence of the periodic boundary conditions. The formation energy was
249
+ initially calculated for different positions of the solute atoms and the configurations
250
+ with lower formation energy was selected as the most stable ones (Yuasa et al., 2014).
251
+ In the binary Mg47N1 (N = Y, Ca) alloys, the most stable configuration was found when
252
+ one Mg atom at the center site of the stacking fault plane was substituted by a solute
253
+ atom X. In the ternary Mg46N1X1 (N = Y, and X = Ca) alloy, the most stable
254
+ configuration was found when one Mg atom at the center site of the stacking fault plane
255
+ was substituted by a Ca atom. Then, one of the eleven nearest Mg atoms from the Ca
256
+ atom was substituted by one Y atom, as shown in Fig. S1 in the supplementary material.
257
+ The exact position of the Y atom was determined from the formation energy (Ding et
258
+ al., 2019; Dong et al., 2018). The formation energies for every occupancy of the Y atom
259
+ are listed in Table S1 in the supplementary material.
260
+ The conventional direct crystal slip methods were employed to obtain the GSFE
261
+ curves of different slip systems The perfect supercell was cut into two free parts and
262
+
263
+
264
+ 9
265
+ one part was displaced with respect to the other one along the slip direction. The atomic
266
+ positions were relaxed only along the direction perpendicular to the stacking fault plane
267
+ (Wang et al., 2020). A residual force threshold of 0.01 eV/Å was performed in all
268
+ geometric relaxations until the electronic energy converged to less than 10-5 eV/cell.
269
+ The Brillouin zone for the GSFE of the basal slip system, the prismatic slip system, and
270
+ the pyramidal slip system was set as 8×8×1, 10×6×1, and 6×10×1, respectively, with
271
+ an energy cutoff of 480 eV (Dong et al., 2018; Wang et al., 2013).
272
+
273
+
274
+ Fig. 1. Schematic illustration of the models to calculate the GSFE for (a) basal slip (b)
275
+ prismatic slip, and (c) pyramidal Ⅰ slip. The most stable positions of Y and Ca atoms
276
+ determined by the lowest formation energy are marked by blue and purple atoms,
277
+ respectively. Stacking fault planes are noted by the dotted lines.
278
+
279
+ 3. Results
280
+ 3.1 Mechanical behavior of polycrystals
281
+ The inverse pole figure (IPF) map of the as-extruded Mg-Y-Ca alloy along the
282
+ (a) basal slip
283
+ (b) prismatic slip
284
+ (c) pyramidalⅠslip
285
+ [11!00]
286
+ [0001]
287
+ [112!0]
288
+ [101!1]
289
+ [112!0]
290
+ [112!3]
291
+ 105°
292
+ [112!0]
293
+ [11!00]
294
+ [0001]
295
+ Mg
296
+ Ca
297
+ Y
298
+
299
+ :
300
+ O
301
+ O
302
+ O
303
+ O
304
+ 10
305
+ extrusion direction is plotted in Fig. 2a. The {0001} pole figure shows that the Mg-Y-
306
+ Ca alloy possesses a weak texture with a strength of ~ 8.21 mrd, as displayed in Fig.
307
+ 2b, compared to pure wrought Mg with a strong basal texture of >15 mrd (Yin et al.,
308
+ 2021). The engineering stress-strain curves of the extruded Mg-Y-Ca alloy from the
309
+ tensile and compressive tests parallel to the extrusion direction are plotted in Fig. 2c.
310
+ The scatter was very limited and the average tensile elongation was very large (≈ 32%).
311
+ Moreover, the tensile yield stress was 104 MPa, very close to the yield strength in the
312
+ compression tests (122 MPa). Thus, the Mg-Y-Ca alloy presented very low
313
+ tension/compression asymmetry in the yield strength in contrast with the marked
314
+ asymmetry in extruded Mg and Mg alloys (Sukedai and Yokoyama, 2010; Yin et al.,
315
+ 2021; Zhang et al., 2016b).1 It should also be noted that volume fraction of the twinned
316
+ material after tensile deformation was very low (≈ 1.8%), indicating that twining was
317
+ not a dominant deformation mechanism in the Mg alloy.
318
+
319
+ Fig. 2. (a) IPF map of the Mg-Y-Ca along the extrusion direction. (b) {0001} Pole figure
320
+ of the Mg-Y-Ca alloy illustrating the texture characteristics before the deformation in
321
+ the TD-ED plane. (c) Engineering stress-strain curves in tension and compression
322
+ parallel to the extrusion direction of the Mg-Y-Ca alloy.
323
+
324
+ 3.2 Deformation mechanisms
325
+
326
+ 1 The comparison between both curves shows the limited tension-compression anisotropy in the yield
327
+ strength but the differences in the elastic and fully plastic regions are due to the limitations of the
328
+ compression tests. Compression tests always underestimate the elastic modulus because it is very difficult
329
+ to ensure that the specimen surface and the loading plate surface are perfectly parallel. Thus, partial
330
+ contact between both surface leads to localized plastic deformation and to an apparent elastic modulus
331
+ that is lower than the real one. Moreover, barreling of the cylindrical specimen during compression leads
332
+ to non-homogeneous plastic deformation and overestimates the strain hardening for large plastic strains.
333
+ 50μm
334
+ TD
335
+ ED
336
+ Max=8.21
337
+ ED∥ Tensile direction
338
+ 8.21
339
+ 0.00
340
+ (a)
341
+ (b)
342
+ (c)
343
+ (c)
344
+
345
+ 400
346
+ Tension
347
+ Compression
348
+ 300
349
+ 200
350
+ 0
351
+ 10
352
+ 20
353
+ 30
354
+ 40
355
+ Engineering strain (%)Tscedan-(0001) -Magnesium
356
+ 8.21
357
+ 则量计数:100008
358
+ Subset1
359
+ 半宽:10.0*
360
+ 样品对称性:三料
361
+ 使用样本对疗性:数量
362
+ 投射类型:等围积
363
+ 透射平面:XY
364
+ 率球:上
365
+ 00'0
366
+ 11
367
+ The IPF map with the crystallographic orientation of the grains in the Mg-Y-Ca
368
+ alloy is depicted in Fig. 3. The grains were larger than 150 μm, and the micropillars
369
+ were milled from the center of the grains to ensure that they were single crystals. Four
370
+ grains with appropriate orientations (Fig. 3) were selected to activate different
371
+ deformation mechanisms. The loading directions in the four grains are listed in Table 1,
372
+ as well as the maximum Schmid Factor (SF) for the corresponding slip systems (<a>
373
+ basal slip, <a> prismatic slip, <a> pyramidal Ⅰ slip, <c+a> pyramidal Ⅰ slip and <c+a>
374
+ pyramidal Ⅱ slip) as well as {101$2} tensile twinning. The inclination angle in Table 1
375
+ indicates the angle between the c-axis of each grain and the compression direction, as
376
+ presented. The compression direction is nearly parallel to [112$0], [101$0], and [0001] in
377
+ grains B, C and D, respectively, and forms an angle of ~ 48.5° with respect to [0001]
378
+ axis in grain A. Herein, grain A presents the highest SF for <a> basal slip, which is
379
+ prone to be the dominant deformation mechanism during compression. Plastic
380
+ deformation along the <c+a> pyramidal I and II systems is favored in Grain D. <a>
381
+ prismatic and pyramidal as well as <c+a> pyramidal slip systems have similar SFs in
382
+ grain B, while grain C is suitably oriented to promote tensile twinning and <a>
383
+ prismatic slip.
384
+
385
+ Table 1. The loading direction, inclination angle, elastic modulus, and maximum
386
+ Schmid factor for each slip system and tensile twinning in the selected grains.
387
+ Grain
388
+ Loading
389
+ direction
390
+ Inclination
391
+ angle (°)
392
+ Elastic
393
+ modulus
394
+ (GPa)
395
+ Maximum Schmid factor
396
+ Basal
397
+ <a>
398
+ Prismatic
399
+ <a>
400
+ Pyramidal
401
+ Ⅰ <a>
402
+ Pyramidal
403
+ Ⅰ <c+a>
404
+ Pyramidal
405
+ Ⅱ <c+a>
406
+ Tensile
407
+ twin
408
+ A
409
+ [112!3]
410
+ 48.5
411
+ 46.04
412
+ 0.44
413
+ 0.25
414
+ 0.42
415
+ 0.36
416
+ 0.20
417
+ 0.17
418
+ B
419
+ [112!0]
420
+ 83.4
421
+ 48.14
422
+ 0.11
423
+ 0.48
424
+ 0.46
425
+ 0.47
426
+ 0.47
427
+ 0.43
428
+ C
429
+ [101!0]
430
+ 87.8
431
+ 46.48
432
+ 0.03
433
+ 0.46
434
+ 0.42
435
+ 0.43
436
+ 0.37
437
+ 0.49
438
+ D
439
+ [0001]
440
+ 4.5
441
+ 47.47
442
+ 0.06
443
+ 0.00
444
+ 0.03
445
+ 0.44
446
+ 0.47
447
+ -*
448
+ *Tensile twinning cannot be activated during compression along the c-axis.
449
+
450
+
451
+ 12
452
+
453
+ Fig. 3. Inverse pole figure (IPF) map showing the crystallographic orientation of the
454
+ grains in the Mg-Y-Ca alloy. (a) The loading direction in micropillars from grains A and
455
+ B form an angle of ~ 48.5° with the [0001] crystal orientation and are parallel to [112$0],
456
+ respectively. (b) The loading direction in micropillars from grains C is parallel to [101$0].
457
+ (c) The loading direction in micropillars from grain D is parallel to [0001]. The
458
+ compression loading direction is perpendicular to the paper.
459
+
460
+ 3.2.1 Deformation mechanisms in micropillar of grain A
461
+ The engineering stress-strain curves obtained from the compression micropillars
462
+ carved from grain A along [112$3] orientation are plotted in Fig. 4a. For the sake of
463
+ clarity, the horizontal axis of the green curve in Fig. 4a is shifted by 0.5%. After the
464
+ initial elastic region, the curves show gradual yielding and reach a plateau in the flow
465
+ stress at an applied strain of ~ 5%, without significant work hardening afterwards. This
466
+ behavior is consistent with a plastic deformation dominated by basal slip in pure Mg
467
+ and Mg alloys (Kiener et al., 2021; Y. Liu et al., 2017; Luo et al., 2022; Wang et al.,
468
+ 2020, 2019a; Wu et al., 2020). Small strain bursts (noticed by sudden drops in the stress)
469
+ are present in the stress-strain curves and they are associated with the activation of
470
+ dislocation sources in particular basal slip planes. However, the magnitude of the strain
471
+ 400μm
472
+ 200μm
473
+ 011!0
474
+ 0001
475
+ 1!21!0
476
+ Loading direction
477
+ 200μm
478
+ Grain A
479
+ Grain B
480
+ Grain C
481
+ Grain D
482
+ (a)
483
+ (b)
484
+ (c)
485
+
486
+
487
+ 13
488
+ bursts is much smaller than that reported in other Mg alloys. In fact, large strain bursts
489
+ are associated with the localization of deformation in a few slip planes along the
490
+ micropillar (Wang et al., 2019). However, the lateral and top views of the micropillar
491
+ after deformation (Figs. 4c and 4d, respectively) show evidence of uniform slip traces
492
+ along the length and width of the micropillar, indicating that plastic deformation was
493
+ homogeneous. A yield stress of 65 ± 11 MPa (indicated by the black stars in the inset
494
+ of Fig. 4a) was determined from the critical points in the engineering stress-strain
495
+ curves when the curves deviated from linearity, following the procedure detailed in
496
+ (Alizadeh and LLorca, 2020; Wang et al., 2019a).
497
+ Secondary electron images of lateral and top views of the deformed micropillars
498
+ were obtained in the SEM to ascertain the actual deformation mechanisms and are
499
+ shown in Figs. 4c and 4d, respectively. Many parallel slip traces appear on the top and
500
+ lateral surfaces, which were not present before deformation (Fig. 4b). The orientation
501
+ of the slip traces on the micropillar surfaces is indicated by the green dashed lines in
502
+ Figs. 4c and 4d. The slip steps were obviously observed on the top view from the top
503
+ right corner to the lower left corner, and the corresponding slip direction is determined
504
+ as marked with a white arrow in Fig. 4d. They are indicated by blue planes and red
505
+ arrows, respectively, in Figs. 4e and 4f within the crystallographic lattice. It is evident
506
+ that the slip traces in the micropillar are parallel to the basal planes and the shear
507
+ deformation takes place along the [21$1$0] direction, as shown from the top and lateral
508
+ views of the deformed micropillar. In fact, the (0001) <21$1$0> basal slip system has
509
+ highest SF (listed in Table 1) and plastic deformation along this slip system is dominant
510
+ in this micropillar. Therefore, the CRSS for <a> basal slip (based on the yield stress and
511
+ the corresponding SF) can be estimated as 29 ± 5 MPa.
512
+
513
+
514
+ 14
515
+
516
+ Fig. 4. (a) Engineering stress-strain curves obtained from micropillar compression tests
517
+ in grain A. The yield stress is marked with a black star. SEM images of the micropillar
518
+ (b) before compression and after compression from the (c) right lateral view and (d) top
519
+ view. The slip plane trace and slip direction are indicated by the green dashed lines and
520
+ the white arrow, respectively. The schematic crystallographic lattice of the
521
+ corresponding slip plane is presented (e) for right side and (f) for top side. The blue
522
+ planes indicate the theoretical basal glide planes, and the red arrows represent the
523
+ corresponding shear directions.
524
+
525
+ 3.2.2 Deformation mechanisms in micropillars of grains B and C
526
+ Representative engineering stress-strain curves obtained from micropillar
527
+
528
+ 150
529
+ Engineering stress (MPa)
530
+ Top side
531
+ 100
532
+ :0
533
+ 50
534
+ Right
535
+ side
536
+ 0
537
+ 2um
538
+ 3
539
+ 6
540
+ 9
541
+ 12
542
+ Engineering strain (%)
543
+ Right side
544
+ Top side
545
+ Slip
546
+ Tr.Basal plane
547
+ direction
548
+ Tr.Basal plane
549
+ 2μm
550
+ 2μm
551
+ Basal slip plane
552
+ Basal slip plane
553
+ 15
554
+ compression tests along [112$0] in grain B and along [101$0] in grain C are plotted in
555
+ Figs. 5a and 5b, respectively. The horizontal axis of the green and blue curves was
556
+ shifted by -0.1% and +0.1%, respectively, in the inset of Fig. 5b for the sake of clarity.
557
+ The stress-strain curves are smooth, without distinct strain bursts. The initial elastic
558
+ region is followed by another linear plastic region with reduced strain hardening rate.
559
+ This behavior is radically different from that observed in micropillars with equivalent
560
+ orientation in pure Mg and several Mg alloys (Mg-Al, Mg-Zn, Mg-Y and Mg-Zn-Ca)
561
+ (Li et al., 2021b, 2021a; Y. Liu et al., 2017; Wang et al. 2021a, 2020; Wu et al., 2020),
562
+ which presented large strain bursts after the initial elastic region due to the nucleation
563
+ of tensile twins at the top of the micropillar. They are similar to those found in Mg-2Y
564
+ (wt. %) alloy at 250 ℃ (Li et al., 2021b), where <a> prismatic slip replaced twinning
565
+ as the dominant plastic deformation mechanisms. The yield stresses (obtained as
566
+ indicated above and marked with purple stars in Fig. 5) were 219 ± 9 MPa and 228 ±
567
+ 4 MPa along [112$0] and [101$0] orientations, respectively.
568
+
569
+ Fig. 5. (a) Engineering stress-strain curves obtained from micropillar compression tests
570
+ in grain B along [112$0]. (b) Idem in grain C along [101$0].
571
+
572
+ The representative morphology of the micropillar deformed along [112$0] (grain B)
573
+ is depicted in the SEM images in Figs. 6a and 6b from two different sides (front and
574
+ left, respectively). Faint slip traces are visible on both lateral surfaces of the deformed
575
+ micropillars, as indicated by the blue dashed lines in Figs. 6c and 6d, which show the
576
+ rectangular zones marked by dashed lines in Figs. 6a and 6b, respectively, at higher
577
+ magnification. The slip traces are distributed homogeneously along the lateral surfaces,
578
+ Grain B:[112!0]
579
+ (a)
580
+ (b)
581
+ Grain C:[101!0]
582
+
583
+ 350
584
+ 300
585
+ 250
586
+ 200
587
+ 50
588
+ 100
589
+ 50
590
+ 0
591
+ 2
592
+ 4
593
+ 6
594
+ 8
595
+ 10
596
+ 12
597
+ 0
598
+ Engineering strain (%)240
599
+ 220
600
+ 200
601
+ 0.8
602
+ 1.3
603
+ 1.8280
604
+ 230
605
+ 180
606
+ 1.2
607
+ 1.7
608
+ 2.2350
609
+ Grain C:10101
610
+ 300
611
+ D
612
+ 250
613
+ stress
614
+ 200
615
+ 150
616
+ 100
617
+ Engineering
618
+ strain
619
+ 16
620
+ indicating that plastic deformation was uniform along the micropillar. Moreover, there
621
+ are not slip steps at the surface (as opposed to the micropillar deformed along [112$3] in
622
+ Figs. 4c and 4d), in agreement with the smooth stress-strain curves. This deformation
623
+ morphology is different from that observed in other Mg alloys compressed along a-axis
624
+ (Li et al., 2021b, 2021a; Y. Liu et al., 2017; Wang et al., 2020; Wu et al., 2020), where
625
+ two regions with different contrast were always observed after the deformation due to
626
+ the nucleation of tensile twins.
627
+
628
+ (a)
629
+ 2μm
630
+ Front side
631
+ Left side
632
+ 2μm
633
+ (c)
634
+ (b)
635
+ 1μm
636
+ Tr. Prismatic
637
+ plane
638
+ (d)
639
+ 1μm
640
+ (h)
641
+ (i)
642
+ Prismatic slip plane
643
+ Left side
644
+ Front side
645
+ 2μm
646
+ 4
647
+ 0
648
+ 2μm
649
+ 2μm
650
+ (e)
651
+ (f)
652
+ (g)
653
+ Tr. Prismatic
654
+ plane
655
+ Before
656
+ deformation
657
+ After
658
+ deformation
659
+
660
+ KAM
661
+ Kernel Aver. Misorient.
662
+ 0
663
+ [o]
664
+ 3.96
665
+ Y1
666
+ 5μm
667
+ 光栅:498x331
668
+ 步长尺寸:0.025μm
669
+ >X12μm
670
+ 光栅:243x1692
671
+ 17
672
+ Fig. 6. SEM images of the micropillar deformed along [112$0] from grain B. (a) Lateral
673
+ front and (b) lateral left view side. The traces of the active slip planes are indicated by
674
+ the blue dashed lines in Figs. 5c and 5d, which show the rectangular zones marked by
675
+ dashed lines in Figs. 5a and 5b, respectively, at higher magnification. (e) and (f) TKD
676
+ maps of the lamella extracted from the undeformed region in grain B and along the
677
+ compression direction from the deformed micropillar, respectively. (g) KAM map of
678
+ the deformed micropillar. (h) and (i) Schematics of the crystallographic lattice showing
679
+ the corresponding slip plane for the lateral front side and left side, respectively. The red
680
+ planes indicate the theoretical prismatic glide planes, and the blue lines represent the
681
+ corresponding slip traces.
682
+
683
+ In order to identify the deformation mechanisms, two parallel thin foils were
684
+ extracted from the undeformed region in grain B and along the loading direction from
685
+ the deformed micropillar, respectively, and their orientation was determined by TKD.
686
+ The position of the lamellae is indicated in Fig. S2 of the supplementary material. The
687
+ corresponding orientation maps in Figs. 6e and 6f show that the IPF map (∥Z) of the
688
+ undeformed and deformed thin foils share the same orientation and demonstrate that
689
+ tensile twins were not nucleated during micropillar compression up to 10% strain.
690
+ Moreover, Fig. 6g presents the kernel average misorientation (KAM) map of the whole
691
+ pillar in Fig. 6g reveals the homogeneous deformation without shear bands assuming
692
+ an angular threshold of 4°. The slip traces on the lateral surfaces of the micropillars
693
+ were associated with the prismatic planes, as indicated in Figs. 6h and 6i. Thus,
694
+ prismatic slip was triggered at the onset of the yielding and dominated plastic
695
+ deformation. The maximum SFs for <a> prismatic slip, <a> pyramidal I, <c+a>
696
+ pyramidal II slip and tensile twinning were very similar along this orientation (Table 1)
697
+ but the presence of Y and Ca in solid solution favored the activation of prismatic slip.
698
+ It should be noted that <c+a> pyramidal I slip dominated plastic deformation and
699
+ hindered the development of tensile twinning in micropillar compression tests along
700
+ [1$21$0] orientation in a Mg-4Y (wt. %) (Wu et al., 2020). The maximum SFs for <c+a>
701
+ pyramidal I slip, tensile twinning and <a> prismatic slip in this orientation were 0.41,
702
+ 0.46 and 0.49 and, thus, the preference of <c+a> pyramidal slip can be associated with
703
+ the higher CRSSs for tensile twin nucleation and <a> prismatic slip in Mg-4Y alloy
704
+ (Wu et al., 2020).
705
+
706
+
707
+ 18
708
+ Similar deformation morphology was found in the micropillars deformed along
709
+ [101$0] in grain C. Continuous slip traces were homogeneously distributed along the
710
+ lateral surfaces, as indicated by the dashed blue lines from in Fig. 7b, which shows the
711
+ rectangular region marked by dashed lines in Fig. 7a at higher magnification. As in the
712
+ previous case, the micropillar orientation before and after deformation was assessed by
713
+ TKD carried out in a thin lamella extracted from the undeformed region (Fig. 7c) and
714
+ from the deformed micropillar along the loading direction (Fig. 7d), respectively. The
715
+ relative orientation between the two thin lamellas is shown in Fig. S3 in the
716
+ supplementary material. The corresponding orientation maps do not show any evidence
717
+ of tensile twinning and <a> prismatic slip was again the dominant plastic deformation
718
+ mechanism. This conclusion is supported by the uniform plastic deformation without
719
+ obvious shear bands revealed by the KAM map assuming an angular threshold of 2°
720
+ (Fig. 7e) and the agreement between the slip traces on the lateral surfaces with the
721
+ orientation of the prismatic planes in the micropillar (Fig. 7f). Thus, the CRSSs for
722
+ prismatic slip (obtained from the yield stress and the SF for both micropillar
723
+ orientations) were determined to be 105 ± 4 MPa and 105 ± 2 MPa along [112$0] and
724
+ [101$0] orientations, respectively.
725
+
726
+
727
+ 19
728
+
729
+ Fig. 7. SEM images of the micropillar deformed along [101$0] from grain C. (a) Lateral
730
+ left view side and (b) which shows the rectangular zone marked with a dashed line in
731
+ Fig. 7a at higher magnification. The traces of the active slip planes are indicated by the
732
+ blue dashed lines. (c) and (d) TKD maps of the lamella extracted from the undeformed
733
+ region in the grain and along the compression direction from the deformed micropillar,
734
+ respectively. (e) KAM map of the deformed micropillar in grain C. (f) Schematic of the
735
+ crystallographic lattice showing the corresponding slip plane for the lateral left view in
736
+ (a) and (b). The red plane indicates the theoretical prismatic glide plane, and the blue
737
+ 2μm
738
+ (e)
739
+ Prismatic slip plane
740
+ Left side
741
+ 2μm
742
+ (c)
743
+ 1μm
744
+ Left side
745
+ Tr. Prismatic
746
+ plane
747
+ (a)
748
+ (b)
749
+ 2μm
750
+ (d)
751
+ (f)
752
+ 2
753
+ 0
754
+ 2μm
755
+ Before
756
+ deformation
757
+ After
758
+ deformation
759
+
760
+ 2KAM
761
+ Kernel Aver. Misorient.
762
+ 光栅:164x268步长尺寸:0.03μm
763
+ Y1
764
+ 2μm
765
+ ?X1IPF
766
+ IPF Coloring II ZO
767
+ Magnesium
768
+ 0001
769
+ -12-10
770
+ 01-10
771
+ Y1
772
+ 2μm
773
+ 光栅:220x145
774
+ 步长尺寸:0.04μm
775
+ >X1
776
+ 20
777
+ line represents the corresponding traces.
778
+
779
+ Further assessment of the deformation mechanisms was carried out by means of
780
+ TEM observations of the dislocation structures in a thin lamella extracted from the
781
+ micropillar deformed along [101$0] (Fig. 8). The lamella was nearly parallel to (1$21$0)
782
+ plane, as confirmed by the SADP in the inset in Fig. 8a, and there are no traces of
783
+ twinning in the micropillar. Dark field micrographs of the square region marked in Fig.
784
+ 8a are depicted in Figs. 8b and 8c with g = (101$0) and g = (0002), respectively. Large
785
+ density of dislocations is observed in Fig. 8b but they disappear from this region when
786
+ g = (0002) in Fig. 8c. They are obviously <a> dislocations with 1/3 a [112$0] or 1/3
787
+ [21$1$0] Burgers vector, based on the dislocation extinguish condition. However, the SF
788
+ of the {011$0} [21$1$0] prismatic slip system is very low (~0.05), thus, it is reasonable to
789
+ assume that the Burgers vector of the <a> dislocations in Fig. 8b is 1/3 [112$0]. The <a>
790
+ screw dislocations are observed under g = (101$0) condition as marked with yellow
791
+ arrows in Fig. 8b. The Burgers vector of screw dislocation is parallel to the dislocation
792
+ line, leading to the straight dislocation lines nearly parallel to trace of the basal planes
793
+ (marked with a green line).
794
+
795
+
796
+ 21
797
+
798
+ Fig. 8. TEM micrographs of the lamella extracted from the micropillar deformed along
799
+ [101$0]. The beam direction is parallel to [1$21$0] orientation. (a) Low magnification view
800
+ of the lamella. (b) and (c) High magnification dark field micrographs with g = (101$0)
801
+ and g = (0002), respectively, from the region marked with a blue square in (a).
802
+
803
+ The activation of the <a> prismatic slip during compression along the a-axis has
804
+ been reported recently in Mg-Zn-Ca alloy (Wang et al., 2021b) in combination with
805
+ tensile twinning. However, the activation of <a> prismatic slip and the suppression of
806
+ tensile twinning during compression along the a-axis has not been found at ambient
807
+ temperature in pure Mg (Y. Liu et al., 2017) or any Mg alloys (Li et al., 2021b, 2021a;
808
+ Wang et al., 2021a, 2020; Wu et al., 2020). This result is very surprising because
809
+ compression of Mg and its alloys along the a-axis (or equivalent extension along the c-
810
+ axis) easily leads to the nucleation and growth of {101$2} tensile twins, because the
811
+ associated CRSS to promote tensile twin is much lower than that necessary to activate
812
+ 5μm
813
+ (a)
814
+ 200nm
815
+ 200nm
816
+ (b)
817
+ (c)
818
+ g=(101!0)
819
+ g=(0002)
820
+ <a> dislocations
821
+ (0001) plane
822
+ B=[1!21!0]
823
+
824
+
825
+ 22
826
+ <c+a> pyramidal slip or <a> prismatic slip. While the addition of Y and Ca in solid
827
+ solution leads to a large increase in the CRSS for <a> prismatic slip with respect to pure
828
+ Mg (from 39 MPa in pure Mg (Kaya, 2013) to 105 MPa), it seems to have a much larger
829
+ effect on the CRSS for twin nucleation. In fact, considering the maximum stresses
830
+ attained in the micropillar compression tests along [112$0] and [101$0] orientations (258
831
+ MPa in Fig. 5a and 303 MPa in Fig. 5b, respectively) and the maximum SFs for tensile
832
+ twinning in both orientations (Table 1), it can be estimated that the CRSS for twin
833
+ nucleation in the Mg-Y-Ca alloys should be higher than 148 MPa.
834
+
835
+ 3.2.3 Deformation mechanisms in micropillar of grain D
836
+ The engineering stress-strain curves obtained from the compression micropillars
837
+ carved from grain D along [0001] orientation are plotted in Fig. 9a. After the elastic
838
+ region, a strong linear hardening was observed in the plastic region. The yield stress
839
+ (marked by the purple stars in the inset) was 431 ± 15 MPa. This mechanical response
840
+ is in good agreement with the results reported in Mg-0.4Y (wt. %) and Mg-4Y (wt. %)
841
+ alloys (Wu et al., 2020) as well as in precipitation-hardened Mg-4Zn (wt. %) alloy
842
+ (Alizadeh et al., 2021) under c-axis compression. In all these cases, the presence of Y
843
+ in solid solution or of β1
844
+ ' precipitates increased the CRSS for basal slip and plastic
845
+ deformation was accommodated through <c+a> pyramidal slip due to the low SF of
846
+ basal planes in this orientation. The SEM micrograph of the lateral side of deformed
847
+ micropillar in Fig. 9b, shows no slip traces but this behavior is also typical of pyramidal
848
+ slip, which does not lead to visible slip traces on the micropillar surface. The orientation
849
+ of the micropillar after deformation was assessed by TKD in a thin lamella extracted
850
+ along the compression direction. The IPF map in Fig. 9c indicates the absence of the
851
+ tensile twinning during deformation.
852
+
853
+
854
+ 23
855
+
856
+ Fig. 9. (a) Engineering stress-strain curves from the micropillar deformed in
857
+ compression along [0001] in grain D. (b) SEM micropillar of the lateral side of the
858
+ deformed micropillar. (c) IPF map of the lamella extracted from the deformed
859
+ micropillar.
860
+
861
+ To further elucidate the deformation mechanisms, the analysis of the dislocation
862
+ structures was carried out by TEM in a thin lamella extracted from the deformed
863
+ micropillar. The beam direction was parallel to [112$0] as confirmed by the SADP in the
864
+ inset in Fig. 10a. Two-beam condition imaging was performed with g = (0002) and g =
865
+ (101$0) and the corresponding dark field micrographs are depicted in Figs. 10b and Fig.
866
+ 10c, respectively. A large density of <c+a> dislocations (marked with blue arrows) is
867
+ observed under g = (0002) in Fig. 10b, and some <a> components are still in contrast
868
+ at the same location when the operation vector changes to g = (101$0) in Fig. 10c. The
869
+ detail of the rectangular region marked with purple dashed lines in Fig. 10b is shown at
870
+ higher magnification in Fig. 10d. The <c+a> dislocations (marked with the blue dashed
871
+ lines) are [1$1$23]/3 and [12$1$3]/3 according to the [112$0] crystal orientation in Fig. 10e.
872
+ These results are in agreement with those reported in Mg-Zn-Ca and Mg-Y alloys
873
+ (Wang et al., 2021a; Wu et al., 2020). Nevertheless, it should be noticed that it is
874
+ difficult to identify the active pyramidal plane, since both pyramidal I and pyramidal II
875
+ planes contain the same slip directions. Thus, it can be concluded that plastic
876
+ deformation along the c-axis in compression was dominated by <c+a> pyramidal
877
+ dislocations. The activation of pyramidal slip was associated to homogeneous
878
+ deformation and strong strain hardening (Basu et al., 2021). This high hardening rate is
879
+ likely associated with short mean-free paths and this explains why no slip traces were
880
+ 2μm
881
+ (b)
882
+ After compression
883
+ (a)
884
+ 1μm
885
+ (c)
886
+
887
+ 2700
888
+ 600
889
+ 90
890
+ 500
891
+ 400
892
+ 300
893
+ 200
894
+ 100
895
+ 10
896
+ ineeiring
897
+ Stran500
898
+ 400
899
+ 300
900
+ 2
901
+ 3.T
902
+ 24
903
+ found on the micropillar surface. It is not clear whether slip took place along pyramidal
904
+ I plane or pyramidal II plane but the SF is slightly higher for pyramidal II for this
905
+ particular orientation (Table 1), which is assumed to be the active one. Thus, the CRSS
906
+ for <c+a> pyramidal Ⅱ slip can be estimated as 203 ± 7 MPa from the SF and the yield
907
+ stress.
908
+
909
+ Fig. 10. (a) Bright field TEM image of the thin lamella extracted from the micropillar
910
+ deformed along the [0001] orientation. The beam direction is [112$0], as shown by the
911
+ SADP in the inset. (b) and (c) Dark field TEM micrographs of the square region marked
912
+ with orange dash lines in (a) under g = (0002) and g = (101$0), respectively. (d) Bright
913
+ field TEM micrograph obtained under g = (0002) from the rectangular region marked
914
+ with purple dash lines in (b). The potential <c+a> pyramidal dislocations are indicated
915
+ by the crystal orientation (black box) and the pyramidal slip trace (blue line). (e)
916
+ Schematic of Mg crystal orientation and of the corresponding <c+a> dislocations (blue
917
+ lines) from the [112$0] projected view.
918
+
919
+ 3.3 Effect of Y and Ca on the GSFE curves
920
+ (a)
921
+ 1μm
922
+ g=(101!0)
923
+ <a> components
924
+ (0001) plane
925
+ <c+a> dislocation
926
+ 100nm
927
+ 100nm
928
+ (b)
929
+ (c)
930
+ (d)
931
+ (e)
932
+ 50nm
933
+ g=(0002)
934
+ B=[112!0]
935
+ g=(0002)
936
+
937
+ Basal trace
938
+ 25
939
+ The experimental evidence presented above shows that the addition of Y and Ca
940
+ affects significantly the plastic deformation mechanisms. The changes in the
941
+ deformation mechanisms are proposed to be associated with the modification of the slip
942
+ resistance of the different slip systems due to the presence of the solute atoms. The
943
+ GSFE is intimately associated with the activation barriers of the deformation modes,
944
+ hence influencing their relative contributions to the overall deformation behavior
945
+ (Sandlöbes et al., 2011). To ascertain the effect of the solute atoms (Y and/or Ca) on the
946
+ slip activities, the GSFE (γ) curves were computed for the <a> slip systems in Mg-Y,
947
+ Mg-Ca, and Mg-Y-Ca alloy, as well as in pure Mg for comparison.
948
+ The GSFE curves for {0001}<101$0>, {11$00}<112$0> and {101$1}<1$21$0> slip
949
+ systems are presented in Figs. 11a, 11b and 11c, respectively. The curves exhibited only
950
+ one local maximum, from which the unstable stacking fault energy (γus) for each slip
951
+ system was determined (Table 2). γus is associated with the activation barrier for
952
+ dislocation slip (Ding et al., 2018; Dong et al., 2018). Evidently, the addition of Y and
953
+ Ca reduced slightly the γus for <a> basal slip from 88 mJ/m2 in pure Mg to a minimum
954
+ of 64 mJ/m2 in Mg-Ca or of 73 mJ/m2 in Mg-Y and the γus of Mg-Y-Ca (74 mJ/m2) was
955
+ similar with that of Mg-Y. However, the reduction in γus for the <a> prismatic slip
956
+ system was much more important, from ~235 mJ/m2 in pure Mg to γus of ~18 mJ/m2 in
957
+ the Mg-Y-Ca alloy (Table 2). This synergistic contribution of Y and Ca on γus for <a>
958
+ prismatic slip is obvious as the sole addition of either Y or Ca only reduced γus to 120
959
+ mJ/m2 (Table 2). The dramatic reduction of γus for <a> prismatic slip in the Mg-Y-Ca
960
+ alloy facilitates the activation of this deformation mechanism during plastic
961
+ deformation. On the contrary, the γus for <a> pyramidal Ⅰ only changed from 304 mJ/m2
962
+ in pure Mg to 318 mJ/m2 in Mg-Y-Ca alloy. The sole addition of Ca (308 mJ/m2) did
963
+ not modify significantly γus for <a> pyramidal Ⅰ while Y (359 mJ/m2) increased slightly
964
+ γus for <a> pyramidal I. Thus, <a> prismatic slip is favored by the addition of Y and Ca
965
+ in comparison with <a> pyramidal I slip.
966
+
967
+
968
+ 26
969
+
970
+ Fig. 11. Generalized stacking fault energy curves for (a) <a> basal (b) <a> prismatic,
971
+ and (c) <a> pyramidal Ⅰ slip systems in pure Mg, Mg-Ca, Mg-Y and Mg-Y-Ca alloys.
972
+
973
+ Table 2. The calculated γus, for basal <a>, prismatic <a>, and pyramidal Ⅰ <a>, slip
974
+ systems in the Mg-Ca, Mg-Y and Mg-Y-Ca alloys compared with pure Mg.
975
+ Alloy
976
+ γus (mJ/m2)
977
+ <a> Basal
978
+ <a> Prismatic
979
+ <a> Pyramidal Ⅰ
980
+ Mg48
981
+ 88
982
+ 235
983
+ 304
984
+ Mg47Ca1
985
+ 64
986
+ 121
987
+ 308
988
+ Mg47Y1
989
+ 73
990
+ 118
991
+ 359
992
+ Mg46Y1Ca1
993
+ 74
994
+ 18
995
+ 318
996
+
997
+ 4. Discussion
998
+ 4.1 Effect of Y and Ca on the CRSSs
999
+ The yield stresses measured from the micropillar compression tests in different
1000
+ (a)
1001
+ (b)
1002
+ (c)
1003
+
1004
+ ikgas
1005
+ Tas
1006
+ Ipyranniclalsillp
1007
+ 300
1008
+ 20
1009
+ 200
1010
+ 20160
1011
+ 100
1012
+ 60
1013
+ 0.0
1014
+ 0,2)
1015
+ OLA
1016
+ 0.
1017
+ 1.0
1018
+ lraictonaldlspiaxeementof8<112os300
1019
+ pmshatcslp
1020
+ 250
1021
+ 200
1022
+ 160
1023
+ 100
1024
+ (),,
1025
+ 0.6
1026
+ 1.0
1027
+ bracuoneldisplakeementof 1120100
1028
+ lbersalslip
1029
+ 80
1030
+ 6X0
1031
+ 0
1032
+ 0.0
1033
+ 0.)
1034
+ 0.6
1035
+ 0.8
1036
+ 1.0
1037
+ Hracuonealdisplakeementof sslolcs
1038
+ 27
1039
+ orientations are summarized in Table 3. The CRSS for the dominant slip system in each
1040
+ orientation (following slip trace analysis and TEM characterization) is also presented
1041
+ in Table 3. They are <a> basal slip in the micropillars carved from grain A, <a>
1042
+ prismatic slip in the micropillars from grains B and C, and <c+a> pyramidal II slip in
1043
+ the micropillars from grain D. Moreover, twin nucleation was not observed in
1044
+ micropillars carved from grains B and C and this result can be used to obtain thresholds
1045
+ of the CRSS for twin nucleation from the maximum stress attained during the test and
1046
+ the maximum SF for tensile twinning in Table 1. These minimum values are also
1047
+ included in Table 3. It should be noticed that dimensions of the micropillars selected in
1048
+ this investigation follow previous results in Mg alloys (Li et al., 2021a; Wang et al.,
1049
+ 2020; Wu et al., 2020) that indicate these values should not be very much influenced
1050
+ by the “smaller is stronger” effect reported for micropillar compression tests at the
1051
+ micron or sub-micron scale (Aitken et al., 2015; Chang et al., 2014).
1052
+
1053
+ Table 3. Yield stress and CRSS for different slip systems from micropillar compression
1054
+ tests along different orientation in the Mg-Y-Ca alloy.
1055
+ Grain
1056
+ Loading direction
1057
+ Yield stress (MPa)
1058
+ CRSS (MPa)
1059
+ A
1060
+ [112!3]
1061
+ 65 ± 11
1062
+ 29 ± 5 (<a> basal slip)
1063
+ B
1064
+ [112!0]
1065
+ 219 ± 9
1066
+ 105 ± 4 (<a> prismatic slip)
1067
+ > 111 MPa (tensile twin*)
1068
+ C
1069
+ [101!0]
1070
+ 228 ± 4
1071
+ 105 ± 2 (<a> prismatic slip)
1072
+ > 148 MPa (tensile twin*)
1073
+ D
1074
+ [0001]
1075
+ 431 ± 15
1076
+ 203 ± 7 (<c+a> pyramidal Ⅱ slip)
1077
+ *: Tensile twin was not nucleated when the CRSS reached this value.
1078
+
1079
+ In order to ascertain the strengthening effect of Y and Ca atoms in solid solution,
1080
+ the CRSSs for the different slip systems in Mg-Y-Ca alloy are plotted in Fig. 12 along
1081
+ with those reported in the literature in pure Mg (Li et al., 2021a; Wang et al., 2019a),
1082
+ Mg-Al (Wang et al., 2020, 2019a), Mg-Zn (Li, 2019; Li et al., 2021a), Mg-Y (Li et al.,
1083
+ 2021b; Wu et al., 2020), Mg-Zn-Ca (Wang et al., 2021a), Mg-Al-Ca (Luo et al., 2022)
1084
+
1085
+
1086
+ 28
1087
+ and Mg-Y-Zn (Chen et al., 2018) alloys. All these results were obtained from
1088
+ compression tests in micropillars with a cross-section around 5 × 5 μm2 and, thus, size
1089
+ effects -if any- should not affect the comparison. The results for <a> basal slip in Fig.
1090
+ 12a show that the addition of Y in solid solution dramatically increases the CRSS in
1091
+ comparison with pure Mg (Li et al., 2021a; Wang et al., 2019a) and with Mg-Al (Wang
1092
+ et al., 2019a) or Mg-Zn (Li, 2019; Li et al., 2021a) alloys. These experimental data are
1093
+ supported by the first principles simulations of the solute/dislocation interaction energy,
1094
+ which showed the higher strengthening potential of Y for basal dislocations, in
1095
+ comparison with Al and Zn, because of the larger atomic radius and shear modulus
1096
+ misfit of Y with respect to Mg (Tehranchi et al., 2018). The addition of Ca to the Mg-Y
1097
+ does not increase the CRSS for basal slip according to our results while the
1098
+ strengthening effect of Ca in Mg-Al (Luo et al., 2022) or Mg-Zn (Wang et al., 2021a)
1099
+ is limited and may also be attributed to the elastic interaction between Ca solute atoms
1100
+ and dislocations.
1101
+ Regarding <c+a> pyramidal slip (Fig. 12b), Zn and Al are the alloying elements
1102
+ which lead to the largest increase in the CRSS (Li et al., 2021a; Wang et al., 2020). Zn
1103
+ is more efficient but the solubility of Al in Mg is larger and CRSSs in the range of 200-
1104
+ 250 MPa can be achieved for these binary alloys. Addition of 4 wt. % Y increases the
1105
+ CRSS up to 106 MPa (Wu et al., 2020) but the combination of Y and Ca leads to a
1106
+ CRSS of 203 ± 7 MPa, similar to the one found in the binary Mg-Zn alloy. Thus, Zn,
1107
+ Al and Y solutes increase the CRSS for <c+a> pyramidal slip due to the elastic
1108
+ interaction of the solutes with the dislocations, as in the case of <a> basal slip. It should
1109
+ be noticed that the activation and glide of <c+a> pyramidal dislocations is a complex
1110
+ process that also depends on dislocation dissociation during gliding due to the larger
1111
+ Burgers vector (Moitra et al., 2014; Tang and El-Awady, 2014). Atomistic simulations
1112
+ have shown that the presence of Y and Ca favors the activation for cross-slip/double
1113
+ cross-slip of <c+a> pyramidal dislocations, leading to new dislocation loops which can
1114
+ accommodate plastic deformation (Wu et al., 2018).
1115
+
1116
+
1117
+ 29
1118
+
1119
+ Fig. 12. CRSS for (a) basal slip, (b) pyramidal slip and (c) twin nucleation and prismatic
1120
+ slip in Mg and Mg alloys (Chen et al., 2018; Kiener et al., 2021; Li, 2019; Li et al.,
1121
+ 2021b, 2021a; Luo et al., 2022; Wang et al., 2021a, 2020, 2019a; Wu et al., 2020),
1122
+ including the results obtained for Mg-Y-Ca alloy in this investigation. All data were
1123
+ obtained from compression tests in micropillars with a cross-section around 5 × 5 μm2.
1124
+ The arrow in the CRSS for twin nucleation in Mg-Y-Ca indicates that the actual CRSS
1125
+ is higher than the value in the figure.
1126
+
1127
+ The CRSSs for tensile twin nucleation, measured by means of micropillar
1128
+ compression tests in Mg and different Mg alloys, are plotted in Fig. 12c (Kiener et al.,
1129
+ 2021; Li et al., 2021a; Wang et al., 2020, 2021a; Wu et al., 2020). While the addition of
1130
+ Y (Wu et al., 2020) and Al (Wang et al., 2020) lead to the largest enhancements in the
1131
+ CRSS for twin nucleation (the latter because of the larger solid solubility), the highest
1132
+ CRSS is obtained for the ternary Mg-Y-Ca alloy which -following our experimental
1133
+ results- has to be higher than 148 MPa. Generally, the twin nucleation process is
1134
+ dominated by the dislocation-shearing and atomic shuffle. The strong strengthening
1135
+ (a)
1136
+ (b)
1137
+ (c)
1138
+
1139
+ 180
1140
+ 160
1141
+ 140
1142
+ $120
1143
+ 100
1144
+ Prismatic slip
1145
+ hg
1146
+ 410
1147
+ 20
1148
+ 8
1149
+ 10200
1150
+ 150
1151
+ 100
1152
+ nCaamees
1153
+ 10410
1154
+ RSS
1155
+ 10
1156
+ 30
1157
+ provided by Y on the CRSS for twin nucleation can be ascribed to the inhibition of
1158
+ atomic shuffling due to the large atomic radius of Y (0.180 nm). Moreover, Ca has an
1159
+ even larger atomic radius (0.194 nm) and it is proposed that the synergistic contribution
1160
+ of both atoms in solid solution is responsible for the huge increase in the CRSS for twin
1161
+ nucleation. In addition, the elastic interaction of twinning dislocations with different
1162
+ solute atoms also leads to an increase in the CRSS for twin propagation (Ghazisaeidi et
1163
+ al., 2014; Stanford et al., 2015), as it has been reported in previous investigations (Li et
1164
+ al., 2021a, 2021b; Wang et al., 2020, 2021a). However, only the addition of Y and Ca
1165
+ can inhibit twin nucleation in micropillars suitable oriented for twinning, e.g., deformed
1166
+ in compression along [112$0] and [101$0] (Table 1).
1167
+ The high CRSS for tensile twin nucleation in Mg-Y-Ca alloys leads to the
1168
+ activation the <a> prismatic slip, which becomes the dominant plastic deformation
1169
+ mechanism under a-axis compression. There is limited information on the CRSS for <a>
1170
+ prismatic slip (because either <a> basal slip or tensile twinning are usually activated
1171
+ before <a> prismatic slip to accommodate the plastic deformation) and the available
1172
+ experimental data on Mg-Y-Zn (Chen et al., 2018) (102 MPa) and Mg-Y-Ca (105 ± 4
1173
+ MPa) are plotted in Fig. 12c. The CRSS for <a> prismatic slip is much lower than the
1174
+ CRSS for tensile twin nucleation in Mg-Y-Ca and, thus, tensile twinning is suppressed
1175
+ during compression parallel to the a-axis.
1176
+ The CRSSs in Fig. 12 show that the strengthening effect of the Y and Ca for <a>
1177
+ prismatic slip is much lower than the ones reported for <c+a> pyramidal slip and twin
1178
+ nucleation, and also, in relative terms, for <a> basal. Moreover, evidence of <a>
1179
+ prismatic slip is unusual in Mg alloys except in the case of that they contain Ca (Zhu et
1180
+ al., 2019), indicating that the presence of Ca reduces the activation barriers for <a>
1181
+ prismatic slip glide. Besides, Chen et al., (2018) found that <a> prismatic slip was
1182
+ activated during micropillar compression testing of Mg-Y-Zn alloys but it was absent
1183
+ in solution-treated Mg-Zn alloys deformed along the same orientation (Li et al., 2021a;
1184
+ Wang et al., 2019b), implying that the addition of Y also facilitates prismatic slip.
1185
+ Although the elastic interaction of the solute atoms with <a> prismatic dislocations is
1186
+ expected to increase the CRSS, the reduction of the stacking fault energy due to the
1187
+
1188
+
1189
+ 31
1190
+ presence of Y and Ca reduces the activation barrier for dislocation movement on the
1191
+ slip plane and facilitates the activation of this slip system.
1192
+ Overall, the addition of Y and Ca leads to a marked solid solution strengthening
1193
+ for <a> basal and <c+a> pyramidal slip as well as for the nucleation of tensile twins but
1194
+ not for <a> prismatic slip.
1195
+
1196
+ 4.2 Effect of plastic anisotropy on the ductility
1197
+ In general, the tensile ductility and formability of Mg alloys during the plastic
1198
+ deformation is dictated by the CRSS ratio between different slip systems, especially
1199
+ between non-basal and basal slip, the latter being the dominant deformation mechanism
1200
+ in most cases (G. Liu et al., 2017; Zhu et al., 2019). Therefore, the tensile ductility of
1201
+ different Mg alloys is plotted as a function of the CRSS ratios between different slip
1202
+ systems in Fig. 13 (Habibi et al., 2012; Huang et al., 2018; Shi et al., 2020; Wang et al.,
1203
+ 2021b; Wu et al., 2010; Yang et al., 2022; Zhao et al., 2019a, 2019b; Zhu et al., 2020,
1204
+ 2019). The CRSS ratios were measured via micropillar compression tests in most of the
1205
+ alloys (Agnew et al., 2003; Li et al., 2021a; Shang et al., 2021; Wang et al., 2021a, 2020,
1206
+ 2019a, 2018; Wu et al., 2020; Zhu et al., 2019) with a few exceptions. The CRSS ratio
1207
+ between <a> prismatic and <a> basal slip in Mg-5Y (wt. %) (Huang et al., 2018) and
1208
+ Mg-0.47Ca (wt. %) (Zhu et al., 2019) were obtained from mechanical tests in
1209
+ polycrystals via slip trace analysis. Besides, those for pure Mg (Agnew et al., 2003),
1210
+ Mg-0.5Ca (wt. %) (Shang et al., 2021) and Mg-3Y (wt. %) (Wang et al., 2018) alloys
1211
+ were determined by the elasto-plastic self-consistent model, crystal plasticity finite
1212
+ element simulations, and the elastic viscoplastic self-consistent model, respectively.
1213
+ Moreover, the tensile elongation data were collected from pure Mg and wrought Mg
1214
+ alloys with similar grain sizes.
1215
+ In general, reduced ratios between the CRSS for non-basal slip and basal slip are
1216
+ strongly associated with the improvement of the ductility of Mg alloys. This trend
1217
+ agrees with the data plotted in Fig. 13b, which shows a clear link between the reduction
1218
+ of CRSS <c+a> pyramidal / CRSS <a> basal and the increase in tensile elongation. However,
1219
+ the limited data of the influence of CRSS <a> prismatic / CRSS <a> basal on the tensile ductility
1220
+
1221
+
1222
+ 32
1223
+ in Fig. 13a are not conclusive. Obviously, low CRSS <c+a> pyramidal / CRSS <a> basal ratios
1224
+ favor isotropic deformation and limit the development of strong basal textures and both
1225
+ processes help to improve ductility and formability because activation of <c+a>
1226
+ dislocations benefits the strain accommodation along the c-axis (Liu et al., 2019).
1227
+ Besides, Wu et al., (2018) predicted that the addition of Y/Ca could significantly reduce
1228
+ the cross-slip energy barriers between pyramidal I and pyramidal II planes, thus
1229
+ promoting <c+a> dislocation cross-slip. Enhanced non-basal slip activities and cross-
1230
+ slip induce homogeneous deformation and improve the ductility.
1231
+
1232
+ Fig. 13. Relation between the CRSS ratios of different slip systems (Agnew et al., 2003;
1233
+ Li et al., 2021a; Shang et al., 2021; Wang et al., 2021a, 2020, 2019a, 2018; Wu et al.,
1234
+ 2020; Zhu et al., 2019) and the tensile elongation (Habibi et al., 2012; Shi et al., 2020;
1235
+ Wang et al., 2021b; Wu et al., 2010; Yang et al., 2022; Zhao et al., 2019a, 2019b; Zhu
1236
+ et al., 2020, 2019) in pure Mg and Mg alloys: (a) CRSS <a> prismatic / CRSS <a> basal, (b)
1237
+ CRSS <c+a> pyramidal / CRSS <a> basal, (c) CRSS <a> prismatic / CRSS tensile twin.
1238
+
1239
+ It should be noted that these mechanisms are particularly relevant in Mg-Y (Wang
1240
+ (a)
1241
+ (b)
1242
+ (c)
1243
+
1244
+ Pure Mg (Zhu,2020; Agnew,2003)
1245
+ Pure Mg (Habibi, 2012; Agnew,2003)
1246
+ <0.71
1247
+ Mg-0.5Ca(Zhu,2019;Shang,2021)
1248
+ Mg-3Y(Wang,2018;Zhao,2019b)
1249
+ ★Mg-5Y-0.08Ca (This work)
1250
+ 30
1251
+ 10
1252
+ 0.
1253
+ 1.:0
1254
+ 1.8
1255
+ 2.0
1256
+ CRSPure Mg(Li,2021a;Wang.2019a;Habibi,2012)
1257
+ PureMg(LI,2021a;Wang,2019a;Zhu,2020)
1258
+ Mg-5Zn(Shi,2020;Li,2021a)
1259
+ Mg-4Y(Wu,2020;Wu,2010)
1260
+ Mg-4.4Al(Zhao,2019a;Wang,2020)
1261
+ Mg-1.8Zn-0.2Ca (Wang,2021a;Wang,2021b)
1262
+ Mg-5Y-0.08Ca (This work)
1263
+ 20
1264
+ 1030
1265
+ 10
1266
+ Mg-3Y(Wang.2018;Zhao,2019b)
1267
+ Mg-0.47Ca(Zhu,2019)
1268
+ Mg-5Y (Huang,2018;Yang,2020)
1269
+ ★Mg-5Y-0.08Ca (This work)
1270
+ 33
1271
+ et al., 2018) and Mg-Zn-Ca (Wang et al., 2021a, 2021b) alloys as well as in the Mg-Y-
1272
+ Ca alloy analyzed in this investigation. In all these cases, the presence of Y and/or Ca
1273
+ also leads to a high increase in the CRSS for twin nucleation while the CRSS for <a>
1274
+ prismatic slip is not strongly affected. As a result, the CRSS <a> prismatic / CRSS tensile twin
1275
+ is dramatically reduced and this is accompanied by a large increase in the tensile
1276
+ ductility, as shown in Fig. 13c. Particularly, tensile twinning is replaced by <a>
1277
+ prismatic slip during compressive deformation along the a-axis if CRSS <a> prismatic /
1278
+ CRSS tensile twin < 1 and twinning only occurs in grains deformed in tension along the c-
1279
+ axis. Moreover, as the CRSS for <a> prismatic slip is smaller than that for <c+a>
1280
+ pyramidal slip, the former becomes the dominant plastic deformation mechanism in
1281
+ grains suitable oriented for both. It should be noted that <c+a> pyramidal slip is
1282
+ associated with a large strain hardening (Fig. 9a) that it is not present for <a> prismatic
1283
+ slip (Fig. 5). Thus, pyramidal slip induced large stress concentrations at grain
1284
+ boundaries that facilitate the nucleation of damage but this process is not activated if
1285
+ <a> prismatic slip is dominant.
1286
+ In general, the preferential activation of basal slip and tensile twinning during
1287
+ processing always introduces a strong basal texture in wrought Mg and Mg alloys,
1288
+ leading to the plastic anisotropy, crack formation and limited ductility (Sabat et al.,
1289
+ 2015; Wang et al., 2021a). The addition of Y and Ca in our alloy strongly enhanced the
1290
+ activation of prismatic <a> and pyramidal <c+a> slip, which also contribute to reduce
1291
+ the intensity of the texure during extrusion, as shown in Figs. 2a and 2b. This limited
1292
+ texture also contributes to reduce the plastic anisotropy.
1293
+ It should also be noted that the overall mechanical response of polycrystals cannot
1294
+ fully ascertained by means of micromechanical tests in single crystals because other
1295
+ factors (grain boundaries, grain size and texture) play a key role in the mechanical
1296
+ response. However, it should be emphasized that the plastic deformation of each crystal
1297
+ within the polycrystal is intrinsically related to that of a single crystal (Wang et al.,
1298
+ 2021a) and, hence, it is important to ascertain the plastic deformation mechanisms in
1299
+ single crystals to understand the complex mechanisms in bulk polycrystalline samples.
1300
+ Overall, these results indicate that the presence of Y and Ca in solid solution in Mg
1301
+
1302
+
1303
+ 34
1304
+ alloys leads to a large increase in the CRSS for <a> basal slip (which induces a large
1305
+ reduction in CRSS <c+a> pyramidal / CRSS <a> basal) while CRSS <a> prismatic / CRSS tensile twin
1306
+ < 1. As a result, plastic deformation in polycrystals in more isotropic and localization
1307
+ of the deformation in the form intense basal slips that promote fracture is suppressed
1308
+ (Sandlöbes et al., 2011). Moreover, twinning and <c+a> pyramidal slip are replaced by
1309
+ <a> prismatic slip in grains deformed along the a-axis. Suppression of twinning (which
1310
+ induces strong anisotropy in the plastic deformation in textured alloys) and the
1311
+ activation of <a> prismatic slip (which provides an additional plastic deformation
1312
+ mechanism with limited hardening) lead to an important improvement in the tensile
1313
+ ductility of Mg alloys.
1314
+
1315
+ 5. Conclusions
1316
+ The deformation mechanisms of a Mg-5Y-0.08Ca (wt. %) alloy, with a superior
1317
+ tensile elongation (32%), were studied by means of micropillar compression tests, slip
1318
+ trace analysis along different orientations, TEM as well as TKD. It was found that the
1319
+ presence of Y and Ca in solid solution led to a huge increase in the CRSS for <a> basal
1320
+ slip (29 ± 5 MPa), <c+a> pyramidal slip (203 ± 7 MPa) and tensile twin nucleation
1321
+ (above 148 MPa). This behavior was attributed to the large mismatch of the atomic radii
1322
+ and elastic constants of the Y and Ca atoms with respect to Mg, which leads to a strong
1323
+ interaction of the dislocations with the solute atoms and hinders atomic shuffling, that
1324
+ is necessary to activate twin nucleation. On the contrary, the CRSS for <a> prismatic
1325
+ slip only increases up to 105 ± 4 MPa because the hardening induced by the interaction
1326
+ of the solute atoms with dislocations is partially balanced by the reduction in the
1327
+ stacking fault energy associated with <a> prismatic slip due to the presence of Y and
1328
+ Ca.
1329
+ The changes in the CRSS for slip and tensile twinning in Mg-Y-Ca alloys modify
1330
+ the dominant deformation mechanisms. In particular, the CRSS <a> prismatic / CRSS tensile
1331
+ twin is dramatically reduced and tensile twinning is replaced by <a> prismatic slip during
1332
+ compressive deformation along the a-axis if CRSS <a> prismatic / CRSS tensile twin < 1.
1333
+ Moreover, as the CRSS for <a> prismatic slip is smaller than that for <c+a> pyramidal
1334
+
1335
+
1336
+ 35
1337
+ slip, the former becomes the dominant plastic deformation mechanism in grains suitable
1338
+ oriented for both. As a result, reduction of twinning (which induces strong anisotropy
1339
+ in the plastic deformation in textured alloys) and the activation of <a> prismatic slip
1340
+ (which provides an additional plastic deformation mechanism with limited hardening)
1341
+ lead to an important improvement in the tensile ductility of Mg alloys.
1342
+
1343
+ Acknowledgements
1344
+ This work was supported by the National Natural Science Foundation of China
1345
+ (Grant Nos. 52001199 and 51825101). Y. Cui acknowledges the support from the
1346
+ Shanghai Sailing Program (Grant No. 22YF1419300). JLL acknowledges the support
1347
+ from the Spanish Ministry of Science (HexaGB project, reference RTI2018-098245)
1348
+ and from the MAT4.0-CM project funded by the Comunidad de Madrid under
1349
+ programme S2018/NMT-4381.
1350
+
1351
+ References
1352
+ Agnew, S.R., Duygulu, Ö., 2005. Plastic anisotropy and the role of non-basal slip
1353
+ in magnesium alloy AZ31B. Int. J. Plast. 21, 1161–1193.
1354
+ Agnew, S.R., Tomé, C.N., Brown, D.W., Holden, T.M., Vogel, S.C., 2003. Study
1355
+ of slip mechanisms in a magnesium alloy by neutron diffraction and modeling. Scr.
1356
+ Mater. 48, 1003–1008.
1357
+ Ahmad, R., Yin, B., Wu, Z., Curtin, W.A., 2019. Designing high ductility in
1358
+ magnesium alloys. Acta Mater. 172, 161–184.
1359
+ Aitken, Z.H., Fan, H., El-Awady, J.A., Greer, J.R., 2015. The effect of size,
1360
+ orientation and alloying on the deformation of AZ31 nanopillars. J. Mech. Phys. Solids
1361
+ 76, 208–223.
1362
+ Alizadeh, R., LLorca, J., 2020. Interactions between basal dislocations and β1′
1363
+ precipitates in Mg–4Zn alloy: Mechanisms and strengthening. Acta Mater. 186, 475–
1364
+ 486.
1365
+ Alizadeh, R., Wang, J., LLorca, J., 2021. Precipitate strengthening of pyramidal
1366
+ slip in Mg–Zn alloys. Mater. Sci. Eng. A 804, 140697.
1367
+ Basu, I., Chen, M., Wheeler, J., Schäublin, R.E., Löffler, J.F., 2021. Stacking-fault
1368
+ mediated plasticity and strengthening in lean, rare-earth free magnesium alloys. Acta
1369
+ Mater. 211, 116877.
1370
+ Blöchl, P.E., 1994. Projector augmented-wave method. Phys. Rev. B 50, 17953–
1371
+ 17979.
1372
+ Chang, Y.W., Pozuelo, M., Yang, J.M., 2014. Size-Induced Strengthening in
1373
+ Nanostructured Mg Alloy Micropillars. Mater. Res. Lett. 2, 199–203.
1374
+
1375
+
1376
+ 36
1377
+ Chen, R., Sandlöbes, S., Zehnder, C., Zeng, X., Korte-Kerzel, S., Raabe, D., 2018.
1378
+ Deformation mechanisms, activated slip systems and critical resolved shear stresses in
1379
+ an Mg-LPSO alloy studied by micro-pillar compression. Mater. Des. 154, 203–216.
1380
+ Chun, J.S., Byrne, J.G., 1969. Precipitate strengthening mechanisms in
1381
+ magnesium zinc alloy single crystals. J. Mater. Sci. 4, 861–872.
1382
+ Ding, Z., Liu, W., Sun, H., Li, S., Zhang, D., Zhao, Y., Lavernia, E.J., Zhu, Y.,
1383
+ 2018. Origins and dissociation of pyramidal <c + a> dislocations in magnesium and its
1384
+ alloys. Acta Mater. 146, 265–272.
1385
+ Ding, Z., Zhao, G., Sun, H., Li, S., Ma, F., Lavernia, E.J., Zhu, Y., Liu, W., 2019.
1386
+ Alloying effects on the plasticity of magnesium: comprehensive analysis of influences
1387
+ of all five slip systems. J. Phys. Condens. Matter 32, 015401.
1388
+ Dong, Q., Luo, Z., Zhu, H., Wang, L., Ying, T., Jin, Z., Li, D., Ding, W., Zeng, X.,
1389
+ 2018. Basal-plane stacking-fault energies of Mg alloys: A first-principles study of
1390
+ metallic alloying effects. J. Mater. Sci. Technol. 34, 1773–1780.
1391
+ Fu, W., Wang, R., Wu, K., Kuang, J., Zhang, J., Liu, G., Sun, J., 2019. The
1392
+ influences of multiscale second-phase particles on strength and ductility of cast Mg
1393
+ alloys. J. Mater. Sci. 54, 2628–2647.
1394
+ Ghazisaeidi, M., Hector, L.G., Curtin, W.A., 2014. Solute strengthening of
1395
+ twinning dislocations in Mg alloys. Acta Mater. 80, 278–287.
1396
+ Graff, S., Brocks, W., Steglich, D., 2007. Yielding of magnesium: From single
1397
+ crystal to polycrystalline aggregates. Int. J. Plast. 23, 1957–1978.
1398
+ Habibi, M.K., Pouriayevali, H., Hamouda, A.M.S., Gupta, M., 2012.
1399
+ Differentiating the mechanical response of hybridized Mg nano-composites as a
1400
+ function of strain rate. Mater. Sci. Eng. A 545, 51–60.
1401
+ Hofstetter, J., Rüedi, S., Baumgartner, I., Kilian, H., Mingler, B., Povoden-
1402
+ Karadeniz, E., Pogatscher, S., Uggowitzer, P.J., Löffler, J.F., 2015. Processing and
1403
+ microstructure–property relations of high-strength low-alloy (HSLA) Mg–Zn–Ca
1404
+ alloys. Acta Mater. 98, 423–432.
1405
+ Huang, X., Xin, Y., Cao, Y., Li, W., Huang, G., Zhao, X., Liu, Q., Wu, P., 2022.
1406
+ Understanding the mechanisms of texture evolution in an Mg-2Zn-1Ca alloy during
1407
+ cold rolling and annealing. Int. J. Plast. 158, 103412.
1408
+ Huang, Z., Wang, L., Zhou, B., Fischer, T., Yi, S., Zeng, X., 2018. Observation of
1409
+ non-basal slip in Mg-Y by in situ three-dimensional X-ray diffraction. Scr. Mater. 143,
1410
+ 44–48.
1411
+ Jain, J., Poole, W.J., Sinclair, C.W., Gharghouri, M.A., 2010. Reducing the
1412
+ tension–compression yield asymmetry in a Mg–8Al–0.5Zn alloy via precipitation. Scr.
1413
+ Mater. 62, 301–304.
1414
+ Kaya, A.A., 2013. 2 - Physical metallurgy of magnesium, in: Pekguleryuz, M.O.,
1415
+ Kainer, K.U., Arslan Kaya, A. (Eds.), Fundamentals of Magnesium Alloy Metallurgy.
1416
+ Woodhead Publishing, pp. 33–84.
1417
+ Kiener, D., Jeong, J., Alfreider, M., Konetschnik, R., Oh, S.H., 2021. Prospects of
1418
+ Using Small Scale Testing to Examine Different Deformation Mechanisms in
1419
+ Nanoscale Single Crystals—A Case Study in Mg. Crystals.
1420
+ Kresse, G., Furthmüller, J., 1996. Efficiency of ab-initio total energy calculations
1421
+
1422
+
1423
+ 37
1424
+ for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. 6, 15–
1425
+ 50.
1426
+ Lee, J.-Y., Steglich, D., Lee, M.-G., 2018. Fracture prediction based on a two-
1427
+ surface plasticity law for the anisotropic magnesium alloys AZ31 and ZE10. Int. J. Plast.
1428
+ 105, 1–23.
1429
+ Li, F.-F., Fang, G., 2022. Stress-state dependency of ductile fracture in an extruded
1430
+ magnesium alloy and its underlying mechanisms. Int. J. Plast. 152, 103258.
1431
+ Li, N., 2019. High-throughput screening of strength and creep properties in Mg-
1432
+ Zn alloys.
1433
+ Li, N., Wang, C., Monclús, M.A., Yang, L., Molina-Aldareguia, J.M., 2021a. Solid
1434
+ solution and precipitation strengthening effects in basal slip, extension twinning and
1435
+ pyramidal slip in Mg-Zn alloys. Acta Mater. 221, 117374.
1436
+ Li, N., Yang, L., Wang, C., Monclús, M.A., Shi, D., Molina-Aldareguía, J.M.,
1437
+ 2021b. Deformation mechanisms of basal slip, twinning and non-basal slips in Mg–Y
1438
+ alloy by micropillar compression. Mater. Sci. Eng. A 819, 141408.
1439
+ Liu, B.-Y., Liu, F., Yang, N., Zhai, X.-B., Zhang, L., Yang, Y., Li, B., Li, J., Ma, E.,
1440
+ Nie, J.-F., Shan, Z.-W., 2019. Large plasticity in magnesium mediated by pyramidal
1441
+ dislocations. Science 365, 73–75.
1442
+ Liu, G., Zhang, J., Xi, G., Zuo, R., Liu, S., 2017. Designing Mg alloys with high
1443
+ ductility: Reducing the strength discrepancies between soft deformation modes and
1444
+ hard deformation modes. Acta Mate 141, 1–9.
1445
+ Liu, Y., Li, N., Arul Kumar, M., Pathak, S., Wang, J., McCabe, R.J., Mara, N.A.,
1446
+ Tomé, C.N., 2017. Experimentally quantifying critical stresses associated with basal
1447
+ slip and twinning in magnesium using micropillars. Acta Mate 135, 411–421.
1448
+ Luo, S., Wang, L., Wang, J., Zhu, G., Zeng, X., 2022. Micro-compression of Al2Ca
1449
+ particles in a Mg-Al-Ca alloy. Materialia 21, 101300.
1450
+ Maaß, R., van Petegem, S., Ma, D., Zimmermann, J., Grolimund, D., Roters, F.,
1451
+ van Swygenhoven, H., Raabe, D., 2009. Smaller is stronger: The effect of strain
1452
+ hardening. Acta Mater. 57, 5996–6005.
1453
+ Mayama, T., Noda, M., Chiba, R., Kuroda, M., 2011. Crystal plasticity analysis of
1454
+ texture development in magnesium alloy during extrusion. Int. J. Plast. 27, 1916–1935.
1455
+ Moitra, A., Kim, S.-G., Horstemeyer, M.F., 2014. Solute effect on the <a+c>
1456
+ dislocation nucleation mechanism in magnesium. Acta Mater. 75, 106–112.
1457
+ Momma, K., Izumi, F., 2008. VESTA: a three-dimensional visualization system
1458
+ for electronic and structural analysis. J. Appl. Crystallogr. 41, 653–658.
1459
+ Obara, T., Yoshinga, H., Morozumi, S., 1973. {112$2}<112$3> Slip system in
1460
+ magnesium. Acta Metall. 21, 845–853.
1461
+ Partridge, P.G., 1967. The crystallography and deformation modes of hexagonal
1462
+ close-packed metals. Metall. Rev. 12, 169–194.
1463
+ Reed-Hill, R.E., Robertson, W.D., 1957. Deformation of magnesium single
1464
+ crystals by nonbasal slip. JOM 9, 496–502.
1465
+ Rosalie, J.M., Somekawa, H., Singh, A., Mukai, T., 2012. The effect of size and
1466
+ distribution of rod-shaped β1′ precipitates on the strength and ductility of a Mg–Zn
1467
+ alloy. Mater. Sci. Eng. A 539, 230–237.
1468
+
1469
+
1470
+ 38
1471
+ Sabat, R.K., Mishra, R.K., Sachdev, A.K., Suwas, S., 2015. The deciding role of
1472
+ texture on ductility in a Ce containing Mg alloy. Mater. Lett. 153, 158-161.
1473
+ Sandlöbes, S., Friák, M., Korte-Kerzel, S., Pei, Z., Neugebauer, J., Raabe, D., 2017.
1474
+ A rare-earth free magnesium alloy with improved intrinsic ductility. Sci. Rep. 7, 10458.
1475
+ Sandlöbes, S., Friák, M., Neugebauer, J., Raabe, D., 2013. Basal and non-basal
1476
+ dislocation slip in Mg–Y. Mater. Sci. Eng. A 576, 61–68.
1477
+ Sandlöbes, S., Friák, M., Zaefferer, S., Dick, A., Yi, S., Letzig, D., Pei, Z., Zhu,
1478
+ L.-F., Neugebauer, J., Raabe, D., 2012. The relation between ductility and stacking fault
1479
+ energies in Mg and Mg-Y alloys. Acta Mater. 60, 3011–3021.
1480
+ Sandlöbes, S., Zaefferer, S., Schestakow, I., Yi, S., Gonzalez-Martinez, R., 2011.
1481
+ On the role of non-basal deformation mechanisms for the ductility of Mg and Mg-Y
1482
+ alloys. Acta Mater. 59, 429–439.
1483
+ Shang, X., Zhang, H., Wang, L., Zhu, G., Cui, Z., Fu, M.W., Zeng, X., 2021. The
1484
+ effect of stress state and strain partition mode on the damage behavior of a Mg-Ca alloy.
1485
+ Int. J. Plast. 144, 103040.
1486
+ Shi, Q., Wang, C., Deng, K., Nie, K., Cao, M., Gan, W., Liang, W., 2020. Work
1487
+ hardening and softening behavior of pure Mg influenced by Zn addition investigated
1488
+ via in-situ neutron diffraction. Mater. Sci. Eng. A 772, 138827.
1489
+ Sneddon, I.N., 1965. The relation between load and penetration in the
1490
+ axisymmetric boussinesq problem for a punch of arbitrary profile. Int. J. Eng. Sci. 3,
1491
+ 47–57.
1492
+ Stanford, N., Marceau, R.K.W., Barnett, M.R., 2015. The effect of high yttrium
1493
+ solute concentration on the twinning behaviour of magnesium alloys. Acta Mater. 82,
1494
+ 447–456.
1495
+ Stanford, N., Taylor, A.S., Cizek, P., Siska, F., Ramajayam, M., Barnett, M.R.,
1496
+ 2012. {101¯2} Twinning in magnesium-based lamellar microstructures. Scr. Mater. 67,
1497
+ 704–707.
1498
+ Sukedai, E., Yokoyama, T., 2010. Investigation of tensile–compressive yield
1499
+ asymmetry and the role of deformation twin in extruded pure magnesium. Int. J. Mater.
1500
+ Res. 101, 736–740.
1501
+ Sun, J., Jin, L., Dong, J., Wang, F., Dong, S., Ding, W., Luo, A.A., 2019. Towards
1502
+ high ductility in magnesium alloys - The role of intergranular deformation. Int. J. Plast.
1503
+ 123, 121–132.
1504
+ Tang, D., Zhou, K., Tang, W., Wu, P., Wang, H., 2022. On the inhomogeneous
1505
+ deformation behavior of magnesium alloy beam subjected to bending. Int. J. Plast. 150,
1506
+ 103180.
1507
+ Tang, Y., El-Awady, J.A., 2014. Formation and slip of pyramidal dislocations in
1508
+ hexagonal close-packed magnesium single crystals. Acta Mater. 71, 319–332.
1509
+ Tehranchi, A., Yin, B., Curtin, W.A., 2018. Solute strengthening of basal slip in
1510
+ Mg alloys. Acta Mater. 151, 56–66.
1511
+ Wang, C., Zhang, H.-Y., Wang, H.-Y., Liu, G.-J., Jiang, Q.-C., 2013. Effects of
1512
+ doping atoms on the generalized stacking-fault energies of Mg alloys from first-
1513
+ principles calculations. Scr. Mater. 69, 445–448.
1514
+ Wang, J., Chen, Y., Chen, Z., Llorca, J., Zeng, X., 2021a. Deformation mechanisms
1515
+
1516
+
1517
+ 39
1518
+ of Mg-Ca-Zn alloys studied by means of micropillar compression tests. Acta Mater.
1519
+ 217.
1520
+ Wang, J., Li, N., Alizadeh, R., Monclús, M.A., Cui, Y.W., Molina-Aldareguía, J.M.,
1521
+ LLorca, J., 2019a. Effect of solute content and temperature on the deformation
1522
+ mechanisms and critical resolved shear stress in Mg-Al and Mg-Zn alloys. Acta Mater.
1523
+ 170, 155–165.
1524
+ Wang, J., Molina-Aldareguía, J.M., LLorca, J., 2020. Effect of Al content on the
1525
+ critical resolved shear stress for twin nucleation and growth in Mg alloys. Acta Mater.
1526
+ 188, 215–227.
1527
+ Wang, J., Ramajayam, M., Charrault, E., Stanford, N., 2019b. Quantification of
1528
+ precipitate hardening of twin nucleation and growth in Mg and Mg-5Zn using micro-
1529
+ pillar compression. Acta Mater. 163, 68–77.
1530
+ Wang, J., Stanford, N., 2015. Investigation of precipitate hardening of slip and
1531
+ twinning in Mg5%Zn by micropillar compression. Acta Mater. 100, 53–63.
1532
+ Wang, J., Zhu, G., Wang, L., Vasilev, E., Park, J.-S., Sha, G., Zeng, X., Knezevic,
1533
+ M., 2021b. Origins of high ductility exhibited by an extruded magnesium alloy Mg-
1534
+ 1.8Zn-0.2Ca: Experiments and crystal plasticity modeling. J. Mater. Sci. Technol. 84,
1535
+ 27–42.
1536
+ Wang, L., Huang, Z., Wang, H., Maldar, A., Yi, S., Park, J.S., Kenesei, P.,
1537
+ Lilleodden, E., Zeng, X., 2018. Study of slip activity in a Mg-Y alloy by in situ high
1538
+ energy X-ray diffraction microscopy and elastic viscoplastic self-consistent modeling.
1539
+ Acta Mater. 155, 138–152.
1540
+ Wu, B.L., Zhao, Y.H., Du, X.H., Zhang, Y.D., Wagner, F., Esling, C., 2010.
1541
+ Ductility enhancement of extruded magnesium via yttrium addition. Mater. Sci. Eng. A
1542
+ 527, 4334–4340.
1543
+ Wu, J., Si, S., Takagi, K., Li, T., Mine, Y., Takashima, K., Chiu, Y.L., 2020. Study
1544
+ of basal < a > and pyramidal < c + a > slips in Mg-Y alloys using micro-pillar
1545
+ compression. Philos. Mag. 100, 1454–1475.
1546
+ Wu, Z., Ahmad, R., Yin, B., Sandlöbes, S., Curtin, W.A., 2018. Mechanistic origin
1547
+ and prediction of enhanced ductility in magnesium alloys. Science 359, 447–452.
1548
+ Yaghoobi, M., Chen, Z., Murphy-Leonard, A.D., Sundararaghavan, V., Daly, S.,
1549
+ Allison, J.E., 2022. Deformation twinning and detwinning in extruded Mg-4Al: In-situ
1550
+ experiment and crystal plasticity simulation. Int. J. Plast. 155, 103345.
1551
+ Yang, W., Quan, G.F., Ji, B., Wan, Y.F., Zhou, H., Zheng, J., Yin, D.D., 2022. Effect
1552
+ of Y content and equal channel angular pressing on the microstructure, texture and
1553
+ mechanical property of extruded Mg-Y alloys. J. Magnes. Alloy. 10, 195–208.
1554
+ Yin, D.D., Boehlert, C.J., Long, L.J., Huang, G.H., Zhou, H., Zheng, J., Wang,
1555
+ Q.D., 2021. Tension-compression asymmetry and the underlying slip/twinning activity
1556
+ in extruded Mg–Y sheets. Int. J. Plast. 136, 102878.
1557
+ Yuasa, M., Hayashi, M., Mabuchi, M., Chino, Y., 2014. Improved plastic
1558
+ anisotropy of Mg–Zn–Ca alloys exhibiting high-stretch formability: A first-principles
1559
+ study. Acta Mater. 65, 207–214.
1560
+ Zhang, D., Wen, H., Kumar, M.A., Chen, F., Zhang, L., Beyerlein, I.J., Schoenung,
1561
+ J.M., Mahajan, S., Lavernia, E.J., 2016a. Yield symmetry and reduced strength
1562
+
1563
+
1564
+ 40
1565
+ differential in Mg-2.5Y alloy. Acta Mater. 120, 75–85.
1566
+ Zhang, D., Wen, H., Kumar, M.A., Chen, F., Zhang, L., Beyerlein, I.J., Schoenung,
1567
+ J.M., Mahajan, S., Lavernia, E.J., 2016b. Yield symmetry and reduced strength
1568
+ differential in Mg-2.5Y alloy. Acta Mater. 120, 75–85.
1569
+ Zhang, S., Wang, L., Zhu, G., Diehl, M., Maldar, A., Shang, X., Zeng, X., 2022.
1570
+ Predicting grain boundary damage by machine learning. Int. J. Plast. 150, 103186.
1571
+ Zhao, C., Chen, X., Wang, J., Tu, T., Dai, Y., Shin, K.S., Pan, F., 2019a. Strain
1572
+ Hardening Behavior in Mg–Al Alloys at Room Temperature. Adv. Eng. Mater. 21,
1573
+ 1801062.
1574
+ Zhao, C., Li, Z., Shi, J., Chen, X., Tu, T., Luo, Z., Cheng, R., Atrens, A., Pan, F.,
1575
+ 2019b. Strain hardening behavior of Mg–Y alloys after extrusion process. J. Magnes.
1576
+ Alloy. 7, 672–680.
1577
+ Zhou, N., Zhang, Z., Dong, J., Jin, L., Ding, W., 2013. High ductility of a Mg–Y–
1578
+ Ca alloy via extrusion. Mater. Sci. Eng. A 560, 103–110.
1579
+ Zhu, G., Wang, L., Wang, Jie, Wang, Jian, Park, J.-S., Zeng, X., 2020. Highly
1580
+ deformable Mg–Al–Ca alloy with Al2Ca precipitates. Acta Mater. 200, 236–245.
1581
+ Zhu, G., Wang, L., Zhou, H., Wang, J., Shen, Y., Tu, P., Zhu, H., Liu, W., Jin, P.,
1582
+ Zeng, X., 2019. Improving ductility of a Mg alloy via non-basal <a> slip induced by
1583
+ Ca addition. Int. J. Plast. 120, 164–179.
1584
+
1585
+
9dFLT4oBgHgl3EQfuS8F/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
9dFPT4oBgHgl3EQfYjT_/content/tmp_files/2301.13074v1.pdf.txt ADDED
@@ -0,0 +1,2263 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Prepared for submission to JHEP
2
+ Dynamic Radius Jet Clustering Algorithm
3
+ Biswarup Mukhopadhyayaa, Tousik Samuia, and Ritesh K. Singha
4
+ aDepartment of Physical Sciences, Indian Institute of Science Education and Research Kolkata,
5
+ Mohanpur, 741246, India.
6
7
8
+ Abstract:
9
+ The study of standard QCD jets produced along with fat jets, which may
10
+ appear as a result of the decay of a heavy particle, has become an essential part of collider
11
+ studies.
12
+ Current jet clustering algorithms, which use a fixed radius parameter for the
13
+ formation of jets from the hadrons of an event, may be inadequate to capture the differing
14
+ radius features.
15
+ In this work, we develop an alternative jet clustering algorithm that
16
+ allows the radius to vary dynamically based on local kinematics and distribution in the η-φ
17
+ plane inside each evolving jet. We present the usefulness of this dynamic radius clustering
18
+ algorithm through two Standard Model processes, and thereafter illustrate it for a scenario
19
+ beyond the Standard Model at the 13 TeV LHC.
20
+ arXiv:2301.13074v1 [hep-ph] 30 Jan 2023
21
+
22
+ Contents
23
+ 1
24
+ Introduction
25
+ 1
26
+ 2
27
+ Methodology
28
+ 3
29
+ 2.1
30
+ Standard Sequential Recombination Algorithms
31
+ 3
32
+ 2.2
33
+ Our Proposal: Dynamic Radius Jet Clustering Algorithm
34
+ 4
35
+ 3
36
+ Application to Standard Model Processes
37
+ 7
38
+ 3.1
39
+ Illustration I: pp → tj process
40
+ 8
41
+ 3.2
42
+ Illustration II: pp → V j Subprocess
43
+ 15
44
+ 4
45
+ Usefulness in BSM signals
46
+ 19
47
+ 5
48
+ Summary and Outlook
49
+ 25
50
+ 1
51
+ Introduction
52
+ The physics extraction capacity of any high-energy collider depends crucially on the han-
53
+ dling of coloured particles in various final states. These are produced as partons via ei-
54
+ ther short-distance interactions of quantum chromodynamics (QCD) or electroweak pro-
55
+ cesses [1, 2]. The partons, however, hadronize through long-distance QCD effects which
56
+ are not calculable ab initio. One rather uses semi-empirical methods to predict the prob-
57
+ ability that energetic partons will fragment into more low-energy partons and ultimately
58
+ form colour-neutral hadrons which are observable in the detector. Groups of closely spaced
59
+ hadrons with varied degrees of collimation form ‘jets’ whose identification, isolation, and
60
+ merger are predicted once more with the help of semi-empirical (and by no means uniquely
61
+ decided) algorithms called jet clustering algorithms [3–7]. The aim always remains to define
62
+ jets with such algorithms which most accurately elicit the short-distance physics underly-
63
+ ing the events that are studied. They thus constitute some of our most important tools in
64
+ the analysis of phenomena at colliders.
65
+ In the context of the Large Hadron Collider (LHC), a widely used class of jet criteria is
66
+ based on so-called kt-type sequential recombination jet algorithms [7–13]. These algorithms
67
+ (briefly discussed in the next section) typically try to merge ‘neighbouring’ hadrons to
68
+ identify the group as a jet. The neighbourhood of a hadron is defined by a single radius
69
+ parameter R0 in the η-φ plane of the detector, which is used to quantify the radius (or size)
70
+ of a jet. This is because the hadrons within R0 are merged to form a jet while the hadrons
71
+ outside R0 are not included in that jet. The choices for the value of R0 in these algorithms
72
+ depend on the physics searches one is carrying out. At the 13 TeV LHC, the typical choices
73
+ for R0 are 0.4 or 0.8 for a ‘narrow’ or a ‘fat’ jet, respectively. There are, in addition, jet
74
+ isolation criteria depending on whether one is trying to separate a jet from a hard lepton
75
+ or another hadronic jet. However, the sequential recombination algorithms generally do
76
+ not accommodate varying choices of radii on a jet-by-jet basis in a single event since they
77
+ – 1 –
78
+
79
+ have a single constant parameter that determines the radius of a jet. Separate classifiers
80
+ for a ‘narrow’ jet and a ‘fat’ jet in a single event in the current kt-type algorithms are thus
81
+ difficult to set. An important improvement over the current fixed radius algorithms would
82
+ be to make them adapt the jet radii dynamically jet-by-jet in each event. We make an
83
+ attempt in this direction in this work.
84
+ Our central idea of choosing the radius dynamically of a jet, especially for a boosted fat
85
+ jet, is based on the kinematics of the decay products of the initiating heavy particle. From
86
+ the theoretical side, the formation of boosted fat jet occurs due to the high collimation of
87
+ the on-shell decay products – and their showering and subsequent hadronization – of the
88
+ energetic and therefore boosted heavy particles. This is very different from the formation of
89
+ light quark- or gluon-initiated jets, whose collimation is primarily due to parton showering
90
+ and subsequent hadronization. On the other hand, at the operational level, as per the
91
+ standard kt-type algorithms, the fat jets are formed in the same way as the regular ‘narrow’
92
+ jets, which are initiated by light quarks or gluons. However, the kinematics of on-shell decay
93
+ products and their radiation pattern of a heavy particle is different from the showering of
94
+ energetic light quarks or gluons. Therefore, the internal structure of a fat jet is very different
95
+ from a narrow one. These internal structure has been used to tag different heavy and
96
+ light jets in the LHC context. For example, jet substructure (JSS) observable generalized
97
+ angularities λκ
98
+ β [14, 15] is used to distinguish between quark- and gluon-initiated jets [16–
99
+ 26]. The same variable was used in the classification among the narrow jet, fat W jet,
100
+ or boosted top jet [27–30]. Another important set of JSS observables, namely the energy
101
+ correlation functions (ECFs) [31, 32], was shown to be useful in classifying different types
102
+ of jets [29, 33–37]. The observable N-subjettiness (τN) [38, 39] has been used to find the
103
+ multi-pronged nature of light or heavy jets [40–65]. These variables have also been used
104
+ extensively by the experimental collaborations at the 13 TeV LHC [66–68]. These examples
105
+ try to exploit the energy distribution pattern inside a jet to distinguish a heavy object from
106
+ a QCD jet. The common theme of these jet substructure variables is the utilization of the
107
+ ‘multi-pronged’ nature of the fat jets. Due to this multi-pronged nature, one expects the
108
+ variance of inter-constituent distance ∆R of a fat jet to be significantly different compared
109
+ to the narrow QCD jets.
110
+ This variance of a jet can be used to grow the radius of a
111
+ jet starting from an initial radius. Earlier attempts to make the jet radius variable, albeit
112
+ with somewhat different motivations and formalisms, can be found in references [69, 70]. In
113
+ Ref. [69], the effective radius of a pseudojet during their evolution was taken to be inversely
114
+ proportional to the pT with a maximum cut-off on the radius. Essentially, this algorithm
115
+ starts from a big effective radius and the size shrinks as a process of evolution. On the
116
+ other hand, in Ref. [70], an expectation-maximization approach was taken for clustering
117
+ the hadrons into a pre-determined number of clusters (jets). Our approach, in this work, is
118
+ to modify the standard fixed radius kt-type algorithms to make the radius grow depending
119
+ on the local kinematics and distribution (in the η-φ plane) of the hadrons.
120
+ The rest of the article is organized as follows. In section 2, we briefly outline the kt-type
121
+ sequential recombination algorithms followed by our improvement to the same. We test
122
+ the efficacy of our algorithms on two SM processes and discuss them in section 3. Section 4
123
+ deals with one application in the BSM scenario. We summarize and conclude in section 5.
124
+ – 2 –
125
+
126
+ 2
127
+ Methodology
128
+ 2.1
129
+ Standard Sequential Recombination Algorithms
130
+ At the operational level, a jet is constituted by a bunch of four-momenta obtained using
131
+ some clustering algorithm. Among various possible ways of grouping up the four-momenta
132
+ of an event, we need to choose those relevant to physics at the collider. It is important
133
+ that the clustering algorithm should ensure infrared and collinear (IRC) safety, which, in
134
+ our context, can be defined in terms of the following conditions [7]:
135
+ Infrared (IR) safety: The output of the algorithm should not be affected by the intro-
136
+ duction of a four-momentum with p → 0.
137
+ Collinear (C) safety: The output of the algorithm should not be affected by a collinear
138
+ splitting of any four-momentum.
139
+ The algorithm that best takes care of the issue of IRC safety is known as kt-type sequential
140
+ recombination jet clustering algorithms [7]. We briefly outline these algorithms below1.
141
+ If an event consists of N final state particles, whose four-momenta are taken in a list as
142
+ an input of the kt-type algorithms. The distance dij between the ith and jth four-momenta
143
+ and the distance diB between the ith and the beam are then defined as
144
+ dij = min
145
+
146
+ p2p
147
+ Ti, p2p
148
+ Tj
149
+
150
+ ∆R2
151
+ ij,
152
+ (2.1)
153
+ diB = p2p
154
+ TiR2
155
+ 0,
156
+ (2.2)
157
+ where R0 is the radius parameter of the algorithm, ∆Rij is the Euclidean distance between
158
+ the ith and jth four-momenta in the η-φ plane, and pTi is pT of ith four-momenta. The
159
+ exponent p sets the weight factor to the Euclidean distance in the η-φ plane. The three
160
+ choices of p = 1, 0 and −1 correspond to the kt (KT) [8–10], Cambridge-Aachen (CA) [11,
161
+ 12], and anti-kt (AK) [13] algorithms, respectively. The algorithm for combining nearby
162
+ four-momenta with respect to the above distance measures to form jets has the following
163
+ steps.
164
+ Step 1. The distances dij for all the possible pairs and beam distances diB for all the
165
+ four-momenta are calculated first.
166
+ Step 2. The minimum among all the dij and diB’s is determined.
167
+ Step 3a. If the minimum occurs at one of the i, j pairs, the corresponding ith and jth
168
+ four-momenta are merged to form a new four-momentum. The older ones, ith and
169
+ jth four-momenta are removed from the list and the newly merged one is added to
170
+ the list and goes back to Step 1.
171
+ Step 3b. On the other hand, if the minimum distance is one of the diB, the ith four-
172
+ momenta is declared as a final jet, and it is removed from the list and goes back to
173
+ Step 1.
174
+ 1Here, we only discuss the inclusive algorithms in the LHC context. For other jet clustering algorithms,
175
+ please see Ref. [7].
176
+ – 3 –
177
+
178
+ Step 4. The process is stopped once the list gets empty.
179
+ This class of algorithms is seedless because the clustering of four-momenta to form a
180
+ jet does not start from a particular seed. Rather, the algorithms try to merge the closest
181
+ pair first. A group of hadrons is then declared as a jet when an appropriate size is reached.
182
+ The essential difference among the three different algorithms, viz. AK, CA, and KT is
183
+ that they give different weights to the Euclidean distance in the η-φ plane. This typically
184
+ sets some sort of seed to the clustering algorithms in the sense that it gives a preference
185
+ to a hadron around which four-momenta merge to give rise to a final jet. In the case of
186
+ the KT algorithm, it is the softer (in terms of pT ) constituent which merges first and then
187
+ the harder ones get attached to it. As a result, the shape of the final jet may not be
188
+ circular in the η-φ plane. On the other hand, in the AK algorithm, the hardest particle in
189
+ a neighbourhood becomes some sort of seed for the jet and the softer ones merge at a later
190
+ stage. Hence the final jet looks circular in the η-φ plane. In the CA algorithm, the merging
191
+ is purely angular. Among the three algorithms, the AK algorithm is the most popular one
192
+ owing to its circular shape. Importantly, in the kt-type algorithms, there is a fixed radius
193
+ parameter R0, whose value dictates the typical size of all the jets in a particular event.
194
+ We note that these algorithms are unable to capture the essential features of the events
195
+ where narrow and fat jets may simultaneously arise. In our proposed algorithm, we have
196
+ modified these algorithms to bring out the features of varying sizes of the jets.
197
+ 2.2
198
+ Our Proposal: Dynamic Radius Jet Clustering Algorithm
199
+ The usual kt-type algorithms take a fixed radius as an input parameter, and hence the
200
+ algorithms return all the jets to be of the same size (or narrower) in a single event. This
201
+ lack of dynamicity in choosing a radius can be overcome by setting the radius parameter
202
+ dynamically during the construction of each jet.
203
+ In any kt-type algorithm, the starting point is a list of N four-momenta of particles.
204
+ We will refer to these as fundamental particles or, sometimes, fundamental four-momenta.
205
+ The algorithm follows Steps 1 to 3b, as defined in section 2.1, iteratively until the list
206
+ gets empty. At every iteration, the number of contents of the list gets reduced by one.
207
+ The reduction happens in two ways: (1) via the merger of two four-momenta, (2) via the
208
+ declaration of four-momentum as a final jet. Thus at an intermediate iteration, the list
209
+ contains two different types of objects. These two types of objects are (1) fundamental
210
+ four-momenta, and (2) composite four-momenta, generated through the merger of two or
211
+ more fundamental four-momenta. These composite objects evolve through iterations to
212
+ give rise to the final jets. For our convenience, let us label these composite evolving objects
213
+ as pseudojets. We borrowed the name pseudojet from the PseudoJet class in the FastJet3
214
+ package [71], where all the types of four-momenta are called pseudojet. However, we will
215
+ call them by different names: fundamental, pseudojet (composite or evolving), and jet (or
216
+ final jet).
217
+ Our proposal is to change the constant nature of the radius parameter R0 in Eq. (2.2)
218
+ to a dynamic quantity depending on the distribution, in the η-φ plane, of the fundamental
219
+ objects inside each evolving pseudojet. Therefore, the modified distance measure for the
220
+ – 4 –
221
+
222
+ dynamic radius algorithm takes the form
223
+ dij = min
224
+
225
+ p2p
226
+ Ti, p2p
227
+ Tj
228
+
229
+ ∆R2
230
+ ij,
231
+ (2.3)
232
+ diB = p2p
233
+ Ti R2
234
+ di,
235
+ (2.4)
236
+ where Rdi is the dynamical radius parameter, defined as
237
+ Rdi = R0 + σi.
238
+ (2.5)
239
+ The constant R0 is an input parameter similar to the standard kt-type algorithm and it
240
+ is the starting point of the dynamical growth of the radius of an evolving jet. For the ith
241
+ pseudojet, σi is calculated as
242
+ σ2
243
+ i =
244
+
245
+ a<b
246
+ pTa pTb ∆R2
247
+ ab
248
+
249
+ a<b
250
+ pTa pTb
251
+
252
+
253
+
254
+
255
+
256
+
257
+ a<b
258
+ pTa pTb ∆Rab
259
+
260
+ a<b
261
+ pTa pTb
262
+
263
+
264
+
265
+
266
+ 2
267
+ ,
268
+ (2.6)
269
+ where the summation indices a and b run over the fundamental constituents of the pseudo-
270
+ jet. The modifier σi of the radius parameter in Eq. (2.6) is basically ‘pT -weighted’ standard
271
+ deviation of the distances between pairs of fundamental constituents of an evolving pseu-
272
+ dojet. In our proposal, this standard deviation σi is used to capture the size feature of an
273
+ evolving jet dynamically. For a single fundamental four-momentum, σi is taken to be zero.
274
+ The motivation for choosing the modifier of the radius parameter to be pT -weighted
275
+ standard deviation is as follows. As more than one fundamental objects merge to become
276
+ a new pseudojet, it no longer represents a single point in the η-φ plane; it is a composite
277
+ object whose constituents are distributed in that plane. The standard deviation σi for a
278
+ pseudojet i, defined in Eq. (2.6), provides a measure of its fuzziness. We want to incorporate
279
+ this fuzziness in the radius parameter. In the measure of its fuzziness, we also want the
280
+ harder components to be more dominant than the softer ones. Essentially, if the pseudojet
281
+ is dominated by a single pT -hard fundamental constituent or many extremely collimated
282
+ but similar pT objects, we do not want its radius to get increased further. This is because,
283
+ in these scenarios, the final jet is expected to be a narrow jet.
284
+ On the other hand, if
285
+ the pseudojet has more than one pT -hard fundamental constituents slightly separated, we
286
+ expect it to be a fat jet and therefore need an increment to its radius. Both of these two
287
+ aspects are taken care of by the pT -weighted standard deviation in Eq. (2.6).
288
+ Thus, in our proposal, we first take a starting radius R0 to be our input parameter.
289
+ The algorithm then calculates Rdi for each pseudojet, which at an intermediate state accu-
290
+ mulates some constituents. At every iteration, the value of the dynamic radius parameter
291
+ is calculated as the sum of the starting radius R0 and the radius modifier σ. In a nutshell,
292
+ the proposed algorithm starts from an initial radius R0 and grows its radius dynamically
293
+ using the information from the distribution of its constituents in the η − φ plane.
294
+ In the proposed algorithm, the exponent p to the pT in the expressions of distance
295
+ measures dij and diB in Eqs. (2.3–2.4) can take three possible values. We will call the
296
+ – 5 –
297
+
298
+ corresponding algorithms as dynamic radius AK (DR-AK), dynamic radius CA (DR-CA),
299
+ and dynamic radius KT (DR-KT) jet clustering algorithms.
300
+ The IRC safety of the algorithm through the definitions provided in section 2.1 can
301
+ be approximately ensured in the radius modifier σ as well as in the final output of the
302
+ algorithm. With the introduction of an additional four-momentum, say pq, the additive
303
+ contributions to the numerators and to the denominators of the two terms in Eq. (2.6) can
304
+ be generically written as pTq
305
+
306
+ a
307
+
308
+ pTa∆Rα
309
+ aq
310
+
311
+ (for the denominators, α = 0, and for the two
312
+ numerators α = 1 and 2). Clearly, all the additive contributions go to zero as pTq → 0,
313
+ thereby ensuring the IR safety of the quantity σi for ith pseudojet. For the consideration of
314
+ IR safety of the algorithm, let us assume an extra particle of momentum pq is introduced
315
+ in an existing event. This extra particle actively participates in the clustering process only
316
+ by one of the three actions: (a) by getting merged to another fundamental particle, (b) by
317
+ getting recombined to a composite pseudojet, or (c) by getting declared as a singleton jet.
318
+ The action (a) does not change the value of σ or the four-momentum of the pseudojet after
319
+ the merger of the four-momentum with pq → 0. The same is true for the merger of pq via
320
+ action (b) since the merger of two fundamental four-momenta keeps the value of σ at zero.
321
+ After the merger of this p → 0 four-momentum, both the radius modifier σ and the total
322
+ momentum remain unaffected. After this merger, the rest of the clustering process does
323
+ not get affected, and hence the final output of the clustering algorithm remains unaffected.
324
+ Furthermore, action (c) does not give rise to an extra jet whenever p → 0.
325
+ For the collinear safety, one can see that the radius modifier σ remains almost unaltered
326
+ when a four-momentum is split collinearly. Let a four-momentum pq gets split into pr and
327
+ ps. Any general term pTapTq∆Rα
328
+ aq then becomes pTa(pTr∆Rα
329
+ ar +pTs∆Rα
330
+ as). In the collinear
331
+ splitting limit, pTq = pTr + pTs, ∆Rar = ∆Ras = ∆Raq.
332
+ Moreover, there will not be
333
+ any additional contribution due to the pr and ps combination except for the denominators
334
+ in Eq. (2.6) since ∆Rrs = 0. This ensures an approximate collinear safety of σi for any
335
+ ith pseudojet. On the other hand, for the collinear safety of the algorithm, if any four-
336
+ momentum collinearly splits into two four-momenta, then the distance dij → 0. Hence,
337
+ these two collinearly split four-momenta get merged together at a very early stage; a feature
338
+ that is inherently present in the kt-type algorithm. Other IRC safety features (due to pT -
339
+ dependent prefactors in the dij and diB definitions) of the standard kt-type algorithms will
340
+ be inherited by the dynamic radius algorithm.
341
+ We have implemented the method of dynamic radius jet clustering algorithm as a
342
+ FastJet3 plug-in [71, 72]. This package has many built-in data-types and functionalities
343
+ to optimize the implementation and computation of jet clustering algorithms. In particu-
344
+ lar, we have used NNBase and NNH classes to help us keep track of the distance measures.
345
+ As required by these two classes, our dij measure is also symmetric in i and j indices. The
346
+ ClusterSequence class has then been used to merge two four-momenta and keep track of
347
+ the clustering sequence. The PseudoJet class has been used to store the four-momenta
348
+ information of all the initial, intermediate, and ��nal jets.
349
+ The user info property of
350
+ PseudoJet data-type has been used to store the information related to the radius modifier
351
+ σi of the ith pseudojet. This way of implementation has at most N2 computational com-
352
+ – 6 –
353
+
354
+ plexity for an event of size N. The worst possible complexity arises when all the particles
355
+ in an event are merged to form a single jet. Since this worst possibility does not generally
356
+ occur, we expect the computational expense to be less in a practical scenario. We note that
357
+ the standard kt-type algorithms also have N2 complexity via the basic implementation of
358
+ the FastJet algorithm [71, 72].
359
+ One important point to note is that the equations for distance measures, defined in
360
+ Eqs. (2.1–2.2), can be recast to in the radius parameter in the expression of dij rather than
361
+ in the expression of diB. That is to say that the modified set of equations can be taken to
362
+ be
363
+ ˜
364
+ dij = min
365
+
366
+ p2p
367
+ Ti, p2p
368
+ Tj
369
+ � �∆Rij
370
+ R0
371
+ �2
372
+ (2.7)
373
+ ˜
374
+ diB = p2p
375
+ Ti
376
+ (2.8)
377
+ The standard sequential recombination algorithm yields identical results in both formalisms
378
+ since the radius is a constant parameter. However, if this latter formalism is chosen to
379
+ incorporate dynamicity, the form of the dynamic radius parameter Rd will be different.
380
+ The dynamic radius Rd, in this type of modification, will be dependent on both pseudojets.
381
+ One option would be to add the standard deviations σi and σj of the ith and jth pseudojets,
382
+ respectively, to the constant parameter R0. This way of defining Rd ensures the symmetry
383
+ in i and j indices and, therefore, the implementation of the method via NNBase and NNH
384
+ as a FastJet3 plug-in can easily be performed. In any case, the output of the algorithm
385
+ is modified according to Eq. (2.7–2.8) will be different from that of the one considered in
386
+ Eq. (2.3–2.4).
387
+ We now are ready to apply our formalism to some simple SM processes and check how
388
+ it performs compared to the standard sequential recombination algorithms. We discuss
389
+ this in the next section in connection with SM processes and consider its application to
390
+ BSM in the section after that.
391
+ 3
392
+ Application to Standard Model Processes
393
+ We take the following two SM processes to illustrate the performance of our newly developed
394
+ algorithm.
395
+ I. pp → tj
396
+ II. pp → V j, (V = W or Z)
397
+ For both cases, we have generated parton-level events using MadGraph5 (MG5) [73]. We will
398
+ refer to these events as MG5 parton-level events and the final state partons in these events
399
+ as MG5 partons. A lower cut of 500 GeV on the pT of the jets has been imposed during the
400
+ generation of the MG5 parton-level events. This helps us in generating events with boosted
401
+ top or vector bosons at the parton-level, which then form fat jets after subsequent decays
402
+ and hadronization.
403
+ For the purpose of the following studies, only the hadronic decays
404
+ of top and W/Z are considered. We have then passed the MG5 parton-level events to
405
+ – 7 –
406
+
407
+ Pythia8 [74, 75] for showering and hadronization. The Monash 2013 Tune [76], the default
408
+ tune of Pythia8, has been used to take care of the simulations of underlying events and
409
+ multi-parton interactions in the proton-proton collisions. The output of Pythia8 has then
410
+ been transferred to FastJet3 for the formation of jets.
411
+ 3.1
412
+ Illustration I: pp → tj process
413
+ The top quark, when highly boosted, results in a fat jet while the j yields a narrow jet after
414
+ the effects of showering and hadronization. In order to compare various jet properties be-
415
+ tween the dynamic radius and fixed radius algorithm, we run these two types of algorithms
416
+ on the same set of hadrons from each event. We first demonstrate how the dynamic radius
417
+ algorithms help in capturing the fat and narrow objects in a single event. This has been
418
+ demonstrated by depicting the hadrons and jets of an example event in Fig. 1, where we
419
+ plotted, in the η-φ plane, the position of the hadrons in the event along with the high-pT
420
+ jets constructed out of these hadrons. The sizes of the dots are kept proportional to √pT of
421
+ the hadrons. The jets are represented by the unfilled black circles and the solid dots inside
422
+ the black circles comprise of the constituent hadrons of the jets. The three panels on the
423
+ left show the jets for (a) AK, (c) CA, and (e) KT jet algorithms with R0 = 0.5. In all the
424
+ left panels, the algorithms return three high-pT jets; one near (2,2) position and the other
425
+ two are near (0,5) position in the η-φ plane. With the MG5 parton-level information, we
426
+ identified that the jet in (2,2) position is initiated by j while the two jets near the (0,5)
427
+ position is initiated by the decay products of the hard top quark. Because of fixed radii
428
+ of the standard kt-type jet clustering algorithms, they could not capture all the hadrons
429
+ initiated by the decay products of the top quark inside a single jet; rather they have been
430
+ split into two different jets. A quick fix to this problem would be to increase the size of the
431
+ radius parameter. This prescription, however, ends up increasing the jet size unnecessarily,
432
+ for example at the (2,2) position where such increment is not required. This unnecessary
433
+ increase in the radius of a jet increases jet mass, especially in the high pile-up scenario.
434
+ One interesting option in such cases would be to choose the radius according to the need
435
+ of a jet. This is precisely where the dynamic radius jet clustering algorithm is useful in
436
+ this type of scenario. This can be seen in the three panels on the right in Fig. 1. There the
437
+ hadrons and the high-pT jets are drawn for dynamic radius jet clustering algorithms with
438
+ R0 = 0.5. The interesting point to note in all three right panels is that there are two jets
439
+ instead of three. The radius of the jet near the (0,5) position has been appropriately grown
440
+ to capture the full decay products of the top quark and their radiations while the radius
441
+ of the jet near the (2,2) position did not grow much. This desirable characteristic of a jet
442
+ algorithm would be beneficial for the studies of collider events, where narrow as well as fat
443
+ jets are expected to occur simultaneously. In all the panels, the radius of each black circle
444
+ is kept to be equal to the final radius Rd, as defined in Eq. (2.5), of each individual jet.
445
+ For the fixed radius jet algorithms, the final radius is essentially the fixed radius parameter
446
+ R0.
447
+ Fig. 1 gives an approximate idea of how the dynamic radius helps us in finding a fat jet
448
+ starting from a small radius. Next, we show how often this dynamic radius jet algorithm
449
+ helps us in finding the fat jet. In order to demonstrate that, we have employed the following
450
+ – 8 –
451
+
452
+ −4
453
+ −2
454
+ 0
455
+ 2
456
+ 4
457
+ 0
458
+ 1
459
+ 2
460
+ 3
461
+ 4
462
+ 5
463
+ 6
464
+ φ
465
+ pp → tj
466
+ AK, R0 = 0.5
467
+ (a)
468
+ Hadrons
469
+ −4
470
+ −2
471
+ 0
472
+ 2
473
+ 4
474
+ 0
475
+ 1
476
+ 2
477
+ 3
478
+ 4
479
+ 5
480
+ 6
481
+ φ
482
+ pp → tj
483
+ DR-AK, R0 = 0.5
484
+ (b)
485
+ Hadrons
486
+ −4
487
+ −2
488
+ 0
489
+ 2
490
+ 4
491
+ 0
492
+ 1
493
+ 2
494
+ 3
495
+ 4
496
+ 5
497
+ 6
498
+ φ
499
+ pp → tj
500
+ CA, R0 = 0.5
501
+ (c)
502
+ Hadrons
503
+ −4
504
+ −2
505
+ 0
506
+ 2
507
+ 4
508
+ 0
509
+ 1
510
+ 2
511
+ 3
512
+ 4
513
+ 5
514
+ 6
515
+ φ
516
+ pp → tj
517
+ DR-CA, R0 = 0.5
518
+ (d)
519
+ Hadrons
520
+ −4
521
+ −2
522
+ 0
523
+ 2
524
+ 4
525
+ η
526
+ 0
527
+ 1
528
+ 2
529
+ 3
530
+ 4
531
+ 5
532
+ 6
533
+ φ
534
+ pp → tj
535
+ KT, R0 = 0.5
536
+ (e)
537
+ Hadrons
538
+ −4
539
+ −2
540
+ 0
541
+ 2
542
+ 4
543
+ η
544
+ 0
545
+ 1
546
+ 2
547
+ 3
548
+ 4
549
+ 5
550
+ 6
551
+ φ
552
+ pp → tj
553
+ DR-KT, R0 = 0.5
554
+ (f)
555
+ Hadrons
556
+ Figure 1: Positions of final state hadrons and jets in the η-φ plane in an example event
557
+ for pp → tj process. The red dots represent the final state hadrons and their sizes are kept
558
+ proportional to √pT of the corresponding hadron. The unfilled circles represent the final
559
+ radius Rd of a jet. The teal coloured dots represent the constituents of the hard ‘narrow’
560
+ jets. The green and blue (wherever applicable) dots represent the constituents of the fat
561
+ top jet. The left panel, from top to bottom, is for (a) AK, (c) CA, and (e) KT algorithms
562
+ with R0 = 0.5. The right panel, from top to bottom, represents jets clustered using (b)
563
+ DR-AK, (d) DR-CA, and (f) DR-KT algorithms, respectively, with R0 = 0.5.
564
+ – 9 –
565
+
566
+ procedure. We first form the jets from the hadrons and choose only the high-pT (> 5 GeV)
567
+ jets. We then tag the energetic jets, event by event, as reconstructed ‘top’ or reconstructed
568
+ ‘jet’ with the help of MG5 parton-level information. The events are classified into two
569
+ categories, as described below.
570
+ A1. Category A1 consists of events satisfying the following conditions.
571
+ • A jet should have mass in the range (150, 200) GeV and have ∆R(toptruth, jet) <
572
+ 0.5. This jet is identified as a reconstructed top jet. We label these reconstructed
573
+ objects as ‘top (A1)’ in the subsequent discussions.
574
+ • After the tagging of the top jet, another jet should have pT > 300 GeV and
575
+ should be within 0.5 distance from the original jet as generated by MG5. These
576
+ jets are labelled as ‘jet (A1)’ in further discussions.
577
+ A2. Category A2 are the events which satisfy the following conditions.
578
+ • Two separate jets within 1.0 distance of the original top quark and having
579
+ an invariant mass between 150 and 200 GeV. These two jets are tagged as
580
+ constituent jets of the reconstructed top jet, which is a combination of these
581
+ two constituents. These combinations are labelled as ‘top (A2)’.
582
+ • Another jet having pT > 300 GeV and within 0.5 radius from the original jet.
583
+ This is labelled as ‘jet (A2)’.
584
+ In general, any inclusive kt-type clustering algorithm yields as output many soft jets
585
+ along with the hard ones. The origin of these soft jets is primarily the soft radiation due
586
+ to underlying events and wide angle parton shower. These jets are expected in both the
587
+ category A1 and A2 events. Any jet having pT > 5 GeV and labelled neither as top nor as
588
+ jet is labelled as soft jet.
589
+ The two categories have been chosen to demonstrate the usefulness of the dynamic
590
+ radius jet algorithm. Category A1 captures the whole top jet by the jet clustering algorithm
591
+ while the events in category A2 need post-processing after the jet clustering. Therefore, a
592
+ desirable criterion of a better-performing jet clustering algorithm would be to have more
593
+ events in category A1. In order to illustrate that, for a given category, we define acceptance
594
+ efficiency
595
+ A = number of events accepted in a particular category
596
+ total number of events
597
+ .
598
+ (3.1)
599
+ After the classification of the events into the above two categories, the distribution
600
+ of distances between the MG5 parton-level objects and reconstructed ones are plotted
601
+ in Fig. 2. In both the panels of the figure, the blue and brown histograms are for top
602
+ jets, and the green and red ones are for energetic jets. The corresponding categories of
603
+ the histograms are mentioned alongside the legends. The distributions are shown for jets
604
+ clustered using the AK algorithm with (a) R0 = 0.5, and (b) R0 = 0.8. Since this distance
605
+ between the MG5 parton-level and reconstructed ones are features of parton showering
606
+ and hadronization, the normalized distributions are kind of identical for different radius
607
+ – 10 –
608
+
609
+ 0.0
610
+ 0.1
611
+ 0.2
612
+ 0.3
613
+ 0.4
614
+ 0.5
615
+ ∆R(parton, reconstructed)
616
+ 0
617
+ 2
618
+ 4
619
+ 6
620
+ 8
621
+ 10
622
+ 12
623
+ 14
624
+ frequency (normalized)
625
+ (a)
626
+ AK, R0 = 0.5
627
+ top (A1)
628
+ top (A2)
629
+ jet (A1)
630
+ jet (A2)
631
+ 0.0
632
+ 0.1
633
+ 0.2
634
+ 0.3
635
+ 0.4
636
+ 0.5
637
+ ∆R(parton, reconstructed)
638
+ 0
639
+ 2
640
+ 4
641
+ 6
642
+ 8
643
+ 10
644
+ 12
645
+ 14
646
+ frequency (normalized)
647
+ (b)
648
+ AK, R0 = 0.8
649
+ top (A1)
650
+ top (A2)
651
+ jet (A1)
652
+ jet (A2)
653
+ Figure 2: Normalized distribution of ∆R between the MG5 parton-level object and cor-
654
+ responding reconstructed jet. The jets were clustered using the AK algorithm with radius
655
+ parameters (a) 0.5 and (b) 0.8.
656
+ choices. These ∆R distributions are very similar even with different choices of standard or
657
+ dynamic radius sequential recombination algorithms and, therefore, are not shown to avoid
658
+ repetition. This distribution also justified the choice of 0.5 radius to find reconstructed
659
+ objects from the MG5 partons.
660
+ 500
661
+ 1000
662
+ 1500
663
+ 2000
664
+ 2500
665
+ jet energy [GeV]
666
+ 10−4
667
+ 10−3
668
+ frequency (normalized)
669
+ (a)
670
+ AK, R0 = 0.5
671
+ Category A1
672
+ top
673
+ jet
674
+ 500
675
+ 1000
676
+ 1500
677
+ 2000
678
+ 2500
679
+ jet energy [GeV]
680
+ 10−4
681
+ 10−3
682
+ frequency (normalized)
683
+ (b)
684
+ AK, R0 = 0.5
685
+ Category A2
686
+ top
687
+ jet
688
+ Figure 3: Normalized distribution of jet energy for categories (a) A1 and (b) A2. The
689
+ blue and green histograms are respectively for the reconstructed top and the high-pT jet.
690
+ We show in Fig. 3 the jet energy distributions for the objects of our study. The left
691
+ and right panels show the distributions for categories A1 and A2, respectively. The blue
692
+ and green histograms are for the top and the high-pT jet produced in association with it.
693
+ – 11 –
694
+
695
+ 0
696
+ 50
697
+ 100
698
+ 150
699
+ 200
700
+ jet mass [GeV]
701
+ 0.00
702
+ 0.05
703
+ 0.10
704
+ 0.15
705
+ 0.20
706
+ frequency (normalized)
707
+ (a)
708
+ R0 = 0.5
709
+ Category A1
710
+ DR-AK A = 48.79%
711
+ AK A = 14.71%
712
+ DR-AK top
713
+ DR-AK jet
714
+ DR-AK soft
715
+ AK top
716
+ AK jet
717
+ AK soft
718
+ 0
719
+ 50
720
+ 100
721
+ 150
722
+ 200
723
+ jet mass [GeV]
724
+ 0.00
725
+ 0.05
726
+ 0.10
727
+ 0.15
728
+ 0.20
729
+ frequency (normalized)
730
+ (b)
731
+ R0 = 0.5
732
+ Category A2
733
+ DR-AK A = 20.21%
734
+ AK A = 52.23%
735
+ DR-AK top
736
+ DR-AK jet
737
+ DR-AK soft
738
+ AK top
739
+ AK jet
740
+ AK soft
741
+ 0
742
+ 50
743
+ 100
744
+ 150
745
+ 200
746
+ jet mass [GeV]
747
+ 0.00
748
+ 0.05
749
+ 0.10
750
+ 0.15
751
+ 0.20
752
+ frequency (normalized)
753
+ (c)
754
+ R0 = 0.5
755
+ Category A1
756
+ DR-CA A = 32.52%
757
+ CA A = 14.67%
758
+ DR-CA top
759
+ DR-CA jet
760
+ DR-CA soft
761
+ CA top
762
+ CA jet
763
+ CA soft
764
+ 0
765
+ 50
766
+ 100
767
+ 150
768
+ 200
769
+ jet mass [GeV]
770
+ 0.00
771
+ 0.05
772
+ 0.10
773
+ 0.15
774
+ 0.20
775
+ frequency (normalized)
776
+ (d)
777
+ R0 = 0.5
778
+ Category A2
779
+ DR-CA A = 36.48%
780
+ CA A = 50.43%
781
+ DR-CA top
782
+ DR-CA jet
783
+ DR-CA soft
784
+ CA top
785
+ CA jet
786
+ CA soft
787
+ 0
788
+ 50
789
+ 100
790
+ 150
791
+ 200
792
+ jet mass [GeV]
793
+ 0.00
794
+ 0.05
795
+ 0.10
796
+ 0.15
797
+ 0.20
798
+ frequency (normalized)
799
+ (e)
800
+ R0 = 0.5
801
+ Category A1
802
+ DR-KT A = 38.87%
803
+ KT A = 17.71%
804
+ DR-KT top
805
+ DR-KT jet
806
+ DR-KT soft
807
+ KT top
808
+ KT jet
809
+ KT soft
810
+ 0
811
+ 50
812
+ 100
813
+ 150
814
+ 200
815
+ jet mass [GeV]
816
+ 0.00
817
+ 0.05
818
+ 0.10
819
+ 0.15
820
+ 0.20
821
+ frequency (normalized)
822
+ (f)
823
+ R0 = 0.5
824
+ Category A2
825
+ DR-KT A = 24.70%
826
+ KT A = 47.15%
827
+ DR-KT top
828
+ DR-KT jet
829
+ DR-KT soft
830
+ KT top
831
+ KT jet
832
+ KT soft
833
+ Figure 4: Normalized distribution of jet mass for the process pp → tj. The left panel
834
+ shows the distribution for category A1 events while the right panel is the distribution for
835
+ category A2 events. The blue, green, and red histograms are for reconstructed top, hard
836
+ jet, and soft jets (defined in the text), respectively. The histograms, from top to bottom,
837
+ are for AK, CA, and KT algorithms. The filled histograms correspond to fixed radius
838
+ algorithms and the unfilled ones correspond to their dynamic radius (DR) counterparts.
839
+ – 12 –
840
+
841
+ One of the primary obligations of choosing the appropriate size for jets according to
842
+ requirements is to avoid the rise of jet mass even with soft but widely separated constituents
843
+ inside a jet. We, therefore, choose to show the distribution of masses of reconstructed top
844
+ jets, reconstructed energetic jets in Fig. 4. The jet energy ranges corresponding to the
845
+ mass distributions shown can be approximately 500-2000 GeV, as seen in Fig. 3. The left
846
+ panel of the figure represents the distribution for category A1 events while the right panel
847
+ represents the distribution for category A2 events. The blue, green, and red histograms
848
+ are for reconstructed top, hard jet, and soft jets, respectively. The histograms, from top to
849
+ bottom, are for anti-kt, C/A, and kt algorithms. The filled histograms are for standard jet
850
+ clustering algorithms and the unfilled ones are their dynamic radius counterparts. In the
851
+ legends, the prefix ‘DR’ to AK, CA, or KT stands for dynamic radius. In all the panels, the
852
+ starting radius parameter has been taken to be R0 = 0.5. For standard kt-type algorithms,
853
+ the starting radius is the fixed constant radius parameter, i.e., Rd = R0. The values for
854
+ A for different algorithms and different categories are quoted inside each panel of Fig. 4.
855
+ In all the panels, it is seen that the acceptance efficiencies for A1 category events in the
856
+ cases with dynamic radius algorithms are higher than their fixed radius counterparts.
857
+ An interesting feature to notice is that the mass distribution for the energetic jet re-
858
+ mains almost the same for both the standard and dynamic radius jet clustering algorithms.
859
+ The similarity between these two are more prominent for AK and CA algorithms and less
860
+ so for the KT algorithm. This is expected as the KT algorithm starts to merge softer
861
+ momenta first and then capture the harder ones almost at the end. As a result, this al-
862
+ gorithm lets the size of the dynamic radius grow in the beginning and hence allows the
863
+ softer hadron, even if they are a little wider, to merge with the evolving jet. The top jet
864
+ mass distribution is also a little off with respect to their fixed radius analogue. These are
865
+ not very problematic since jet grooming [77–83], trimming [84], or pruning [85, 86] methods
866
+ help in cleaning soft and wide-angle radiation. A similar strategy of grooming is useful in
867
+ the removal of soft jets as well.
868
+ The change in mass distribution for top jet but not for the energetic jet can easily
869
+ be understood from the behaviour of the final radius Rd = (R0 + σ) [Eq. (2.5)] a jet has
870
+ acquired. We, therefore, show the distribution of the final radii of the three different types
871
+ of jets in Fig. 5. The three plots in the top panel are for category A1 events while those in
872
+ the bottom panel are for category A2 events. For category A2 events, ‘top c1’ and ‘top c2’
873
+ labels represent the two constituent jets of reconstructed top. The distributions are shown
874
+ for DR-AK, DR-CA, and DR-KT algorithms in Figs. 5(a,d), 5(b,e), and 5(c,f), respectively,
875
+ with R0 = 0.5 in each.
876
+ From all the histograms in Fig. 5, some clear features emerge. For the case of category
877
+ A1 top jets, the final radius Rd grows to more than 0.6 with a peak at Rd ≃ 0.75, (ap-
878
+ proximately 50% increase with respect to the starting radius). On the other hand, for the
879
+ energetic jets, Rd does not grow by much. This indicates that the radius grows dynamically
880
+ according to the distribution of constituents inside the jet. The growth of the soft jets is
881
+ higher compared to the hard jets candidates. In general, this is will not be a problem in the
882
+ heavy object finding since they can easily be eliminated by choosing an appropriate pT or
883
+ mass cuts. The story for the category A2 events is similar for jets and soft jets. The only
884
+ – 13 –
885
+
886
+ 0.5
887
+ 0.6
888
+ 0.7
889
+ 0.8
890
+ 0.9
891
+ Rd
892
+ 0
893
+ 5
894
+ 10
895
+ 15
896
+ 20
897
+ frequency (normalized)
898
+ (a)
899
+ Category A1
900
+ DR-AK
901
+ R0 = 0.5
902
+ top
903
+ jet
904
+ soft
905
+ 0.5
906
+ 0.6
907
+ 0.7
908
+ 0.8
909
+ 0.9
910
+ Rd
911
+ 0
912
+ 5
913
+ 10
914
+ 15
915
+ 20
916
+ frequency (normalized)
917
+ (c)
918
+ Category A1
919
+ DR-KT
920
+ R0 = 0.5
921
+ top
922
+ jet
923
+ soft
924
+ 0.5
925
+ 0.6
926
+ 0.7
927
+ 0.8
928
+ 0.9
929
+ Rd
930
+ 0
931
+ 5
932
+ 10
933
+ 15
934
+ 20
935
+ frequency (normalized)
936
+ (b)
937
+ Category A1
938
+ DR-CA
939
+ R0 = 0.5
940
+ top
941
+ jet
942
+ soft
943
+ 0.5
944
+ 0.6
945
+ 0.7
946
+ 0.8
947
+ 0.9
948
+ Rd
949
+ 0
950
+ 5
951
+ 10
952
+ 15
953
+ 20
954
+ frequency (normalized)
955
+ (d)
956
+ Category A2
957
+ DR-AK
958
+ R0 = 0.5
959
+ top c1
960
+ top c2
961
+ jet
962
+ soft
963
+ 0.5
964
+ 0.6
965
+ 0.7
966
+ 0.8
967
+ 0.9
968
+ Rd
969
+ 0
970
+ 5
971
+ 10
972
+ 15
973
+ 20
974
+ frequency (normalized)
975
+ (f)
976
+ Category A2
977
+ DR-KT
978
+ R0 = 0.5
979
+ top c1
980
+ top c2
981
+ jet
982
+ soft
983
+ 0.5
984
+ 0.6
985
+ 0.7
986
+ 0.8
987
+ 0.9
988
+ Rd
989
+ 0
990
+ 5
991
+ 10
992
+ 15
993
+ 20
994
+ frequency (normalized)
995
+ (e)
996
+ Category A2
997
+ DR-CA
998
+ R0 = 0.5
999
+ top c1
1000
+ top c2
1001
+ jet
1002
+ soft
1003
+ Figure 5: Normalized distribution of the final radius Rd of three different types of jets.
1004
+ The three plots in the top panel are for category A1 events while those in the bottom panel
1005
+ are for category A2 events. The conventions for the colours and labels ‘top’, ‘jet’, and ‘soft’
1006
+ are the same as in Fig. 4. For category A2, ‘topc1’ and ‘topc2’ labels represent the two
1007
+ constituent jets of reconstructed top. The distributions are shown for DR-AK, DR-CA,
1008
+ and DR-KT algorithms in the panels (a,d), (b,e), and (e,f), respectively, with R0 = 0.5.
1009
+ difference is that the whole top could not be reconstructed as a single jet in these events.
1010
+ The normalized distributions of the final radii of these two constituent jets of reconstructed
1011
+ tops are plotted. These constituents tend to grow more than the energetic jets.
1012
+ The values of acceptance efficiencies A [Eq. (3.1)] for different category events vary
1013
+ with the choice of the value for the starting radius R0. If the starting radius is small, the
1014
+ algorithms fail to capture the fat jet. On the other hand, the large starting radius R0 will
1015
+ capture the unwanted contamination coming from underlying events or radiations from
1016
+ other nearby showers. As a result, the jets will be unnecessarily fat and massive. There
1017
+ is a suitable range for R0 within which the algorithms work better. We, therefore, show
1018
+ the variation of acceptance efficiencies A as a function of starting radius R0 in Fig. 6 for
1019
+ both categories A1 (blue) and A2 (red). The variations are shown for (DR-) AK, CA, and
1020
+ KT algorithms in panels (a), (b), and (c), respectively. As expected, for small R0 values,
1021
+ the efficiencies for category A1 (blue lines) are negligible in both dynamic radius and fixed
1022
+ – 14 –
1023
+
1024
+ 0.2
1025
+ 0.3
1026
+ 0.4
1027
+ 0.5
1028
+ 0.6
1029
+ 0.7
1030
+ 0.8
1031
+ R0
1032
+ 0
1033
+ 10
1034
+ 20
1035
+ 30
1036
+ 40
1037
+ 50
1038
+ 60
1039
+ A [%]
1040
+ (a)
1041
+ A1, DR-AK
1042
+ A2, DR-AK
1043
+ A1, AK
1044
+ A2, AK
1045
+ 0.2
1046
+ 0.3
1047
+ 0.4
1048
+ 0.5
1049
+ 0.6
1050
+ 0.7
1051
+ 0.8
1052
+ R0
1053
+ 0
1054
+ 10
1055
+ 20
1056
+ 30
1057
+ 40
1058
+ 50
1059
+ 60
1060
+ A [%]
1061
+ (c)
1062
+ A1, DR-KT
1063
+ A2, DR-KT
1064
+ A1, KT
1065
+ A2, KT
1066
+ 0.2
1067
+ 0.3
1068
+ 0.4
1069
+ 0.5
1070
+ 0.6
1071
+ 0.7
1072
+ 0.8
1073
+ R0
1074
+ 0
1075
+ 10
1076
+ 20
1077
+ 30
1078
+ 40
1079
+ 50
1080
+ 60
1081
+ A [%]
1082
+ (b)
1083
+ A1, DR-CA
1084
+ A2, DR-CA
1085
+ A1, CA
1086
+ A2, CA
1087
+ Figure 6: The variation of acceptance efficiency A [Eq. (3.1)] as a function of starting
1088
+ radius R0 for pp → tj SM process. The blue and red lines represent the variations of A for
1089
+ categories A1 and A2 events, respectively. The dashed lines are for (a) AK, (b) CA, and
1090
+ (c) KT algorithm and the solid lines are for their dynamic radius versions.
1091
+ radius analyses since the constituents of the entire top jet could not be captured with these
1092
+ small values of R0. Rather, the category A2 (red lines) which form the top with the help
1093
+ of two jets yields more A . This picture changes once we tend towards higher values for
1094
+ R0 ≃ 0.5 as more and more top jets are being reconstructed in the A1 category. As a result,
1095
+ the values of A for the A2 category get reduced. In all the panels of Fig. 6, it is interesting
1096
+ to note that the dynamic radius algorithms (solid) yield higher values for A than their
1097
+ fixed radius counterparts (dashed). This is indicative of the usefulness of the dynamic
1098
+ radius algorithm over the fixed radius ones.
1099
+ The dip in the blue solid lines after near
1100
+ R0 = 0.7 is not essentially the failure of the algorithm. Rather, it is because of the capture
1101
+ of unwanted contaminations along with the radiation coming from the top. Therefore, the
1102
+ jet mass goes beyond 200 GeV, at which point we stop labelling them as a reconstructed
1103
+ top jet. Furthermore, a rough comparison among the curves in the three panels of Fig. 6
1104
+ indicates that DR-AK is better suited than DR-CA and DR-KT algorithms.
1105
+ 3.2
1106
+ Illustration II: pp → V j Subprocess
1107
+ A similar study has been performed in SM pp → V j, (V = W or Z) processes. In order
1108
+ to ensure the formation of fat jets, a lower cut of 500 GeV on the pT of the jet has been
1109
+ imposed at the time of generation of parton-level events via MG5.
1110
+ These events were
1111
+ then passed on to Pythia8 with Monash 2013 Tune [76] tune for parton showering and
1112
+ hadronization. The final state hadrons of these events were then sent to FastJet3 for jet
1113
+ clustering with starting radius R0 = 0.4.
1114
+ As before, we label the energetic jets coming from a jet clustering algorithm, as recon-
1115
+ structed ‘V’ or reconstructed ‘jet’ with the help of MG5 parton-level information. The rest
1116
+ – 15 –
1117
+
1118
+ 0
1119
+ 50
1120
+ 100
1121
+ jet mass [GeV]
1122
+ 0.00
1123
+ 0.05
1124
+ 0.10
1125
+ 0.15
1126
+ 0.20
1127
+ frequency (normalized)
1128
+ (a)
1129
+ R0 = 0.4
1130
+ Category B1
1131
+ DR-AK A = 72.18%
1132
+ AK A = 61.96%
1133
+ DR-AK V
1134
+ DR-AK jet
1135
+ DR-AK soft
1136
+ AK V
1137
+ AK jet
1138
+ AK soft
1139
+ 0
1140
+ 50
1141
+ 100
1142
+ jet mass [GeV]
1143
+ 0.00
1144
+ 0.05
1145
+ 0.10
1146
+ 0.15
1147
+ 0.20
1148
+ frequency (normalized)
1149
+ (b)
1150
+ R0 = 0.4
1151
+ Category B2
1152
+ DR-AK A = 13.81%
1153
+ AK A = 26.23%
1154
+ DR-AK V
1155
+ DR-AK jet
1156
+ DR-AK soft
1157
+ AK V
1158
+ AK jet
1159
+ AK soft
1160
+ 0
1161
+ 50
1162
+ 100
1163
+ jet mass [GeV]
1164
+ 0.00
1165
+ 0.05
1166
+ 0.10
1167
+ 0.15
1168
+ 0.20
1169
+ frequency (normalized)
1170
+ (c)
1171
+ R0 = 0.4
1172
+ Category B1
1173
+ DR-CA A = 68.46%
1174
+ CA A = 62.30%
1175
+ DR-CA V
1176
+ DR-CA jet
1177
+ DR-CA soft
1178
+ CA V
1179
+ CA jet
1180
+ CA soft
1181
+ 0
1182
+ 50
1183
+ 100
1184
+ jet mass [GeV]
1185
+ 0.00
1186
+ 0.05
1187
+ 0.10
1188
+ 0.15
1189
+ 0.20
1190
+ frequency (normalized)
1191
+ (d)
1192
+ R0 = 0.4
1193
+ Category B2
1194
+ DR-CA A = 18.10%
1195
+ CA A = 25.25%
1196
+ DR-CA V
1197
+ DR-CA jet
1198
+ DR-CA soft
1199
+ CA V
1200
+ CA jet
1201
+ CA soft
1202
+ 0
1203
+ 50
1204
+ 100
1205
+ jet mass [GeV]
1206
+ 0.00
1207
+ 0.05
1208
+ 0.10
1209
+ 0.15
1210
+ 0.20
1211
+ frequency (normalized)
1212
+ (e)
1213
+ R0 = 0.4
1214
+ Category B1
1215
+ DR-KT A = 70.41%
1216
+ KT A = 62.58%
1217
+ DR-KT V
1218
+ DR-KT jet
1219
+ DR-KT soft
1220
+ KT V
1221
+ KT jet
1222
+ KT soft
1223
+ 0
1224
+ 50
1225
+ 100
1226
+ jet mass [GeV]
1227
+ 0.00
1228
+ 0.05
1229
+ 0.10
1230
+ 0.15
1231
+ 0.20
1232
+ frequency (normalized)
1233
+ (f)
1234
+ R0 = 0.4
1235
+ Category B2
1236
+ DR-KT A = 11.99%
1237
+ KT A = 24.02%
1238
+ DR-KT V
1239
+ DR-KT jet
1240
+ DR-KT soft
1241
+ KT V
1242
+ KT jet
1243
+ KT soft
1244
+ Figure 7: Normalized distributions of jet mass for the process pp → V j. The left panel
1245
+ shows jet mass distributions of category B1 events and the right panel is the distribution for
1246
+ category B2 events. The blue, green, and red histograms are for reconstructed V, energetic
1247
+ jet and soft jets (defined in the text), respectively. The histograms, from top to bottom,
1248
+ are for AK, CA, and KT algorithms, respectively. The filled histograms correspond to
1249
+ fixed radius algorithms and the unfilled ones correspond to their dynamic radius (DR)
1250
+ analogues.
1251
+ – 16 –
1252
+
1253
+ of the jets having pT > 5 GeV are tagged as ‘soft jets’. As in the previous illustration, we
1254
+ classify the events into two separate categories based on the following criteria.
1255
+ B1. An event was labelled as a category B1 event if it satisfies the following two conditions.
1256
+ • A jet should have mass in the range (65, 105) GeV and ∆R(VMG5, jet) < 0.5.
1257
+ This jet was identified as a reconstructed V jet and we label them as ‘V (B1)��
1258
+ in further discussions.
1259
+ • After the tagging of the V jet, another jet should have pT > 300 GeV and
1260
+ ∆R(jMG5, jet) < 0.5. These jets are labelled as ‘jet (B1)’ in further discussions.
1261
+ B2. An event, after failing to satisfy the criteria for the category B1, could be classified
1262
+ as a category B2 event subject to satisfying the below conditions.
1263
+ • Two separate jets within 1.0 distance from the original vector boson (W or Z)
1264
+ and should have an invariant mass between 65 and 105 GeV. These two jets are
1265
+ tagged as constituent jets of the reconstructed ‘V’ jet. The final reconstructed
1266
+ ‘V’ jet should be within 0.5 distance from the original boson. This combination
1267
+ is labelled as ‘V (B2)’.
1268
+ • Another jet having pT > 300 GeV and within 0.5 radii of the original jet and
1269
+ this is labelled as ‘jet (B2)’.
1270
+ We show the jet mass distribution in Fig. 7 for SM pp → V j process. All the distribu-
1271
+ tions in the left panel of the figure represent the category B1 events and the distributions
1272
+ in the right panel are for category B2. The blue, green, and red histograms are for recon-
1273
+ structed ‘V’, jet and soft jets, respectively. The histograms, from top to bottom, are for
1274
+ AK, CA, and KT algorithms, respectively. The filled histograms are for standard jet clus-
1275
+ tering algorithms and the unfilled ones are their dynamic radius analogues. Quite clearly,
1276
+ the two peaks in the blue histograms, in all the distributions, correspond to the mass peaks
1277
+ of W and Z bosons. The jet mass distribution of the energetic jets using dynamic radius
1278
+ algorithms remains similar to their fixed radius counterparts. The increment in the per-
1279
+ centage of the acceptance efficiencies A [Eq. (3.1)] of category B1 events is representative
1280
+ of the appropriateness of using the dynamic radius algorithms over the standard ones in
1281
+ these types of scenarios.
1282
+ We next show in Fig. 8 the normalized distributions of the final radius for three different
1283
+ types of jets, viz. ‘V’ jets, energetic jets, and soft jets. As in the pp → tj process, the fat
1284
+ V jets acquires a larger radius than the energetic jets after the dynamical expansion of the
1285
+ jet size. Here, again, the soft jets acquire a higher radius compared to the energetic jets.
1286
+ These soft jetsare not of much concern since they are rather soft and hence they can be
1287
+ removed easily from the analysis.
1288
+ In Fig. 9, we show the variation of A [Eq. (3.1)] as a function of starting radius R0. In
1289
+ all the panels of the figure, the blue and red lines correspond to the variations for categories
1290
+ B1 and B2 events, respectively. The dashed lines are for fixed radius algorithms and the
1291
+ solid lines are for dynamic radius jet algorithms. The variations are shown for (a) AK, (b)
1292
+ – 17 –
1293
+
1294
+ 0.4
1295
+ 0.5
1296
+ 0.6
1297
+ 0.7
1298
+ Rd
1299
+ 0
1300
+ 5
1301
+ 10
1302
+ 15
1303
+ 20
1304
+ frequency (normalized)
1305
+ (a)
1306
+ Category B1
1307
+ DR-AK
1308
+ R0 = 0.4
1309
+ V
1310
+ jet
1311
+ soft
1312
+ 0.4
1313
+ 0.5
1314
+ 0.6
1315
+ 0.7
1316
+ Rd
1317
+ 0
1318
+ 5
1319
+ 10
1320
+ 15
1321
+ 20
1322
+ frequency (normalized)
1323
+ (c)
1324
+ Category B1
1325
+ DR-KT
1326
+ R0 = 0.4
1327
+ V
1328
+ jet
1329
+ soft
1330
+ 0.4
1331
+ 0.5
1332
+ 0.6
1333
+ 0.7
1334
+ Rd
1335
+ 0
1336
+ 5
1337
+ 10
1338
+ 15
1339
+ 20
1340
+ frequency (normalized)
1341
+ (b)
1342
+ Category B1
1343
+ DR-CA
1344
+ R0 = 0.4
1345
+ V
1346
+ jet
1347
+ soft
1348
+ 0.4
1349
+ 0.5
1350
+ 0.6
1351
+ 0.7
1352
+ Rd
1353
+ 0
1354
+ 5
1355
+ 10
1356
+ 15
1357
+ 20
1358
+ frequency (normalized)
1359
+ (d)
1360
+ Category B2
1361
+ DR-AK
1362
+ R0 = 0.4
1363
+ V c1
1364
+ V c2
1365
+ jet
1366
+ soft
1367
+ 0.4
1368
+ 0.5
1369
+ 0.6
1370
+ 0.7
1371
+ Rd
1372
+ 0
1373
+ 5
1374
+ 10
1375
+ 15
1376
+ 20
1377
+ frequency (normalized)
1378
+ (f)
1379
+ Category B2
1380
+ DR-KT
1381
+ R0 = 0.4
1382
+ V c1
1383
+ V c2
1384
+ jet
1385
+ soft
1386
+ 0.4
1387
+ 0.5
1388
+ 0.6
1389
+ 0.7
1390
+ Rd
1391
+ 0
1392
+ 5
1393
+ 10
1394
+ 15
1395
+ 20
1396
+ frequency (normalized)
1397
+ (e)
1398
+ Category B2
1399
+ DR-CA
1400
+ R0 = 0.4
1401
+ V c1
1402
+ V c2
1403
+ jet
1404
+ soft
1405
+ Figure 8: Normalized distribution of final radius Rd for the three different types of jets.
1406
+ The top panel represents the distributions of Rd in the category B1 events and the whole
1407
+ bottom panel is for category B2 events. The conventions for the colours and labels V, jet,
1408
+ and soft are the same as Fig. 7. For category B2 events, ‘V c1’ and ‘V c2’ labels represent
1409
+ the two constituent jets of the reconstructed vector bosons. The distributions are shown for
1410
+ DR-AK, DR-CA, and DR-KT algorithms in the panels (a,d), (b,e), and (c,f), respectively,
1411
+ with R0 = 0.4.
1412
+ CA, and (c) KT algorithms. A quick observation of the curves tells us that the behaviour
1413
+ of these curves is similar to that of the curves in Fig. 6 except the monotonic decreasing
1414
+ nature of the category B2 curves. The reason is as follows: in the case of V jets, the jets
1415
+ are ‘two-pronged’ in nature. Therefore, the small radius jets can capture one of the two
1416
+ prongs of V jets, and thereby these two jets are able to reconstruct V jets in B2 category.
1417
+ However, as the starting radius R0 is increasing, more and more events are migrating to
1418
+ category B1. The declining nature of the curves for large radii after 0.5 is because of the
1419
+ fact that the jets capture more hadrons than are required for their optimal size. As a
1420
+ result, the mass of the V jets tends to go beyond the mass window set to label them as V
1421
+ jets. Again, more variables than just the jet mass can help one to improve the tagger and
1422
+ hence the acceptance efficiency.
1423
+ We conclude this section with the note that the dynamic radius jet algorithms are
1424
+ – 18 –
1425
+
1426
+ 0.2
1427
+ 0.3
1428
+ 0.4
1429
+ 0.5
1430
+ 0.6
1431
+ 0.7
1432
+ 0.8
1433
+ R0
1434
+ 0
1435
+ 10
1436
+ 20
1437
+ 30
1438
+ 40
1439
+ 50
1440
+ 60
1441
+ 70
1442
+ 80
1443
+ Acceptance [%]
1444
+ (a)
1445
+ B1, DR-AK
1446
+ B2, DR-AK
1447
+ B1, AK
1448
+ B2, AK
1449
+ 0.2
1450
+ 0.3
1451
+ 0.4
1452
+ 0.5
1453
+ 0.6
1454
+ 0.7
1455
+ 0.8
1456
+ R0
1457
+ 0
1458
+ 10
1459
+ 20
1460
+ 30
1461
+ 40
1462
+ 50
1463
+ 60
1464
+ 70
1465
+ 80
1466
+ Acceptance [%]
1467
+ (c)
1468
+ B1, DR-KT
1469
+ B2, DR-KT
1470
+ B1, KT
1471
+ B2, KT
1472
+ 0.2
1473
+ 0.3
1474
+ 0.4
1475
+ 0.5
1476
+ 0.6
1477
+ 0.7
1478
+ 0.8
1479
+ R0
1480
+ 0
1481
+ 10
1482
+ 20
1483
+ 30
1484
+ 40
1485
+ 50
1486
+ 60
1487
+ 70
1488
+ 80
1489
+ Acceptance [%]
1490
+ (b)
1491
+ B1, DR-CA
1492
+ B2, DR-CA
1493
+ B1, CA
1494
+ B2, CA
1495
+ Figure 9: The variation of A [Eq. (3.1)] as a function of the starting radius R0 for pp → V j
1496
+ SM process. The blue and red lines represent the values of A for categories B1 and B2
1497
+ events, respectively. The dashed lines are for (a) AK, (b) CA, and (c) KT algorithms. The
1498
+ solid lines are for their dynamic radius versions.
1499
+ useful in finding fat as well as narrow jet in a single event in the colliders.
1500
+ We have
1501
+ successfully illustrated this in two SM processes, viz. pp → tj and pp → V j, at the 13 TeV
1502
+ LHC. A comparison among the three dynamic radius analogues of the standard kt-type
1503
+ algorithm reveals that the DR-AK algorithm performs better compared to the DR-CA or
1504
+ the DR-KT algorithms.
1505
+ 4
1506
+ Usefulness in BSM signals
1507
+ We now illustrate the usefulness of the dynamic radius jet algorithm in the context of
1508
+ a scenario beyond the standard model (BSM). This is a scenario where an additional
1509
+ vectorlike singlet quark b′ of charge −1/3 exists along with (d, s, b). Such quarks occur, for
1510
+ example, in E(6) grand unified theories, as also in some seesaw models of quark masses [87–
1511
+ 92]. The b′ can mix with the three SM down-type quarks when electroweak symmetry
1512
+ breaking takes place2. This causes the mass eigenstate dominated by b′ to decay into a top
1513
+ quark and a W boson. In addition, the mixing between a T3 = −1/2 quark and one with
1514
+ T3 = 0 induces flavour-changing Z- and Higgs-couplings in the b-b′ sector. Thus the b′,
1515
+ produced via strong interactions at the LHC, has the decays b′ → tW, b′ → bZ, b′ → bh.
1516
+ The detailed theoretical framework and the resulting phenomenology have been discussed
1517
+ widely in the literature [59, 93–101].
1518
+ The currently available data from the LHC restrict mb′ to be no less than 1.3–1.5 TeV
1519
+ [102–105]. When such a massive quark decays thereafter, its decay products are consider-
1520
+ 2In the following discussion, we shall (a) denote this mass eigenstate itself by b′, (b) assume that
1521
+ ordinary-exotic quark mixing takes place involving only the third family sequential quark, namely, b, and
1522
+ (c) parametrize the b-b′ mixing by the angle θ.
1523
+ – 19 –
1524
+
1525
+ ably lighter compared to it. Therefore the b′ decay products are considerably boosted, so
1526
+ as to produce fat jets. Furthermore, the difference in mass between two product particles
1527
+ leads to jets of varying degrees of fatness.
1528
+ Since our purpose here is to show the efficacy of the dynamic radius jet algorithm, we
1529
+ illustrate our main points in the context of pp → b′¯b′ followed by each b′ decaying into a
1530
+ top quark and a W boson. The t’s and the W’s thus give rise to energetic jets of different
1531
+ radii. We demonstrate below how our newly developed algorithm can capture the identity
1532
+ of the ensuing final state. While the present work is aimed at capturing the essence of our
1533
+ proposed jet algorithm, a more detailed discussion, including combinations of all the three
1534
+ aforementioned decay channels of the b′, is going to be presented in a separate work [106].
1535
+ mb′
1536
+ sin θL
1537
+ sin θR
1538
+ 1.3 TeV
1539
+ 0.12
1540
+ 8.02 × 10−3
1541
+ Table 1: Values of some important parameters of the vectorlike singlet b′ model considered
1542
+ for the illustration.
1543
+ The model has been implemented in a Mathematica-based package SARAH [107–109].
1544
+ The Universal FeynRules Output (UFO) [110] generated by SARAH is then used in MG5
1545
+ for the generation of parton-level events. The parameter card for MG5 has been generated
1546
+ using spectrum generator SPheno [111, 112]. The values for the important parameters of
1547
+ the model are tabulated in Table 1. The angles θL and θR in the table represent the mixing
1548
+ angle between SM b quark and exotic b′ quark of chirality left and right, respectively. After
1549
+ the generation of the MG5 parton-level events, the rest of the analysis pipeline is the same
1550
+ as the previous illustrations of SM processes.
1551
+ In this illustration, we choose DR-AK, based on the discussion in the previous section.
1552
+ We show the resultant jets having pT > 30 GeV formed out of the hadrons generated by
1553
+ Pythia8 in Fig. 10. The left panel shows the positions of the generated hadrons and jets
1554
+ constructed using the AK algorithm with R0 = 0.5. The right panel shows the same for
1555
+ the DR-AK algorithm. In both panels, the red dots represent the position of final state
1556
+ hadrons in the η-φ plane and the size of each dot is proportional to the √pT of the hadron.
1557
+ The unfilled circles represent the final radius (Rd) of a jet. The teal dots represent the
1558
+ constituents of boosted fat ‘W’ jets. The green, blue, and purple (wherever applicable) dots
1559
+ represent the constituents of the fat ‘top’ jet. The yellow dots containing texts represent
1560
+ the position of the MG5 parton-level pT -hard quarks after the decay of top or W. The
1561
+ mothers of the q or b are mentioned in the subscripts of q or b.
1562
+ An interesting point to observe in Fig. 10(b) is that the DR-AK yields only 4 jets,
1563
+ which are representative of 2 fat W and 2 fat t jets. However, in Fig. 10, the fixed radius
1564
+ algorithm could form the fat W jets but fails to capture the entirety of the two fat t jets.
1565
+ One, of course, can use a bigger radius in the AK algorithm to capture the whole of the
1566
+ top jet. However, this will make the W jet unnecessarily fat. This demonstrates the utility
1567
+ of the dynamic radius jet algorithm.
1568
+ – 20 –
1569
+
1570
+ −4
1571
+ −2
1572
+ 0
1573
+ 2
1574
+ 4
1575
+ η
1576
+ 0
1577
+ 1
1578
+ 2
1579
+ 3
1580
+ 4
1581
+ 5
1582
+ 6
1583
+ φ
1584
+ pp → b′¯b′ → tW −¯tW +
1585
+ AK, R0 = 0.5
1586
+ (a)
1587
+ Hadrons
1588
+ −4
1589
+ −2
1590
+ 0
1591
+ 2
1592
+ 4
1593
+ η
1594
+ 0
1595
+ 1
1596
+ 2
1597
+ 3
1598
+ 4
1599
+ 5
1600
+ 6
1601
+ φ
1602
+ pp → b′¯b′ → tW −¯tW +
1603
+ DR-AK, R0 = 0.5
1604
+ (b)
1605
+ Hadrons
1606
+ Figure 10: The distribution of final state hadrons and jets in η-φ plane for an example
1607
+ event. The colours and sizes of the dots and circles follow the same convention as Fig. 1.
1608
+ The teal coloured dots represent the constituents of hard fat ‘W’ jets. The green and blue
1609
+ (wherever applicable) dots represent the constituents of the fat ‘top’ jet. The yellow dots
1610
+ containing texts represent the position of the hard quarks after the decay of top or W
1611
+ which are mentioned as the subscripts of q or b. The plots are shown for (a) AK and (b)
1612
+ DR-AK algorithms.
1613
+ To study the goodness of DR-AK quantitatively, we define the following criteria for
1614
+ tagging of top and W jets.
1615
+ • A jet having mass in the range (150, 200) GeV and having ∆R(toptruth, jet) < 0.5 is
1616
+ identified as a reconstructed top jet.
1617
+ • A jet will be called W jet if it has a mass in the range (65, 105) GeV and is within
1618
+ 0.5 distance from the original MG5 parton-level W boson.
1619
+ Similar to the illustrations with SM processes, we classify the events into different
1620
+ categories. Due to the complex nature of the final states, we have classified the events into
1621
+ more than two categories in the present scenario. The realization is based on the following
1622
+ understanding.
1623
+ • Out of the two W’s coming directly from b′ in an event, the number of reconstructed
1624
+ W as fat jet from the algorithm could be 0, 1, or 2. We call these reconstructed fat
1625
+ W jets as primary W jets.
1626
+ • Similarly, out of the two t quarks, the number of reconstructed t as fat jets can be 0,
1627
+ 1, or 2.
1628
+ • In some particular cases, the whole top may not be reconstructed, but the W boson
1629
+ coming from the top quarks may be reconstructed. These are referred to as secondary
1630
+ W jets in the subsequent discussions.
1631
+ – 21 –
1632
+
1633
+ Based on the above observations, we classify the events into different categories, whose
1634
+ generic name is given as Cij, where i and j are two integers encoding the number of
1635
+ reconstructed top and reconstructed W’s, respectively. For the present scenario, the allowed
1636
+ value for i does not exceed two. For a given i, the values for j should not exceed 4 − i.
1637
+ That is, to say, i ≤ 2 and j ≤ 4−i. An exhaustive list of all possible categories is tabulated
1638
+ in Table 2. For example, the event shown in Fig. 10 would be categorized as C22 for the
1639
+ DR-AK algorithm while the same event would be classified as C03 for the AK algorithm.
1640
+ One may again subdivide some of the categories into subcategories based on how many W
1641
+ jets are coming directly from b′ (primary W) and how many of them are coming from the
1642
+ decay of the top quark (secondary W). Therefore, the generic name for the subcategories
1643
+ can be given as Cijk with i, j, and k being the numbers of reconstructed top, primary W,
1644
+ and secondary W jets. The possible ranges for i, j, and k are 0 ≤ i, j ≤ 2 and 0 ≤ k ≤ 2−i.
1645
+ Category
1646
+ Subcategory
1647
+ No. of top jet
1648
+ No. of primary
1649
+ No. of secondary
1650
+ W jet
1651
+ W jet
1652
+ C22
1653
+ C220
1654
+ 2
1655
+ 2
1656
+ 0
1657
+ C21
1658
+ C210
1659
+ 2
1660
+ 1
1661
+ 0
1662
+ C20
1663
+ C200
1664
+ 2
1665
+ 0
1666
+ 0
1667
+ C13
1668
+ C121
1669
+ 1
1670
+ 2
1671
+ 1
1672
+ C12
1673
+ C120
1674
+ 1
1675
+ 2
1676
+ 0
1677
+ C111
1678
+ 1
1679
+ 1
1680
+ 1
1681
+ C11
1682
+ C110
1683
+ 1
1684
+ 1
1685
+ 0
1686
+ C101
1687
+ 1
1688
+ 0
1689
+ 1
1690
+ C10
1691
+ C100
1692
+ 1
1693
+ 0
1694
+ 0
1695
+ C04
1696
+ C022
1697
+ 0
1698
+ 2
1699
+ 2
1700
+ C03
1701
+ C021
1702
+ 0
1703
+ 2
1704
+ 1
1705
+ C012
1706
+ 0
1707
+ 1
1708
+ 2
1709
+ C02
1710
+ C020
1711
+ 0
1712
+ 2
1713
+ 0
1714
+ C011
1715
+ 0
1716
+ 1
1717
+ 1
1718
+ C002
1719
+ 0
1720
+ 0
1721
+ 2
1722
+ C01
1723
+ C010
1724
+ 0
1725
+ 1
1726
+ 0
1727
+ C001
1728
+ 0
1729
+ 0
1730
+ 1
1731
+ C00
1732
+ C000
1733
+ 0
1734
+ 0
1735
+ 0
1736
+ Table 2: The definitions of the list of categories and subcategories as according to how
1737
+ many fat jets can be reconstructed from the jet algorithm.
1738
+ – 22 –
1739
+
1740
+ 0
1741
+ 50
1742
+ 100
1743
+ 150
1744
+ 200
1745
+ jet mass [GeV]
1746
+ 0.00
1747
+ 0.05
1748
+ 0.10
1749
+ 0.15
1750
+ 0.20
1751
+ frequency (normalized)
1752
+ (a)
1753
+ R0 = 0.5
1754
+ Category C22
1755
+ DR-AK A = 5.47%
1756
+ AK A = 1.62%
1757
+ DR-AK top
1758
+ DR-AK W
1759
+ DR-AK soft
1760
+ AK top
1761
+ AK W
1762
+ AK soft
1763
+ 0.5
1764
+ 0.6
1765
+ 0.7
1766
+ 0.8
1767
+ 0.9
1768
+ Rd
1769
+ 0
1770
+ 5
1771
+ 10
1772
+ 15
1773
+ frequency (normalized)
1774
+ (b)
1775
+ Category C22
1776
+ DR-AK
1777
+ top
1778
+ W
1779
+ soft
1780
+ Figure 11: (a) The normalized distribution of jet mass of the category C22 events for the
1781
+ pp → b′¯b′ → tW −¯tW + process. The blue, green, and red histograms are reconstructed top,
1782
+ W, and soft jets, respectively. The unfilled histograms are for the jets clustered using the
1783
+ DR-AK algorithm while the filled ones are for the jets using the AK clustering algorithm.
1784
+ (b) The normalized distribution of the final radii of top, W, and soft jets with blue, green,
1785
+ and red colours, respectively. For both panels, R0 = 0.5 was used and any additional jets
1786
+ having pT >5 GeV were considered as a soft jet.
1787
+ We plot the normalized distribution of jet mass of the category C22 events in Fig. 11(a).
1788
+ Jets were clustered using R0 = 0.5. In the plot, the blue, green, and red histograms are
1789
+ reconstructed top, W, and soft jets, respectively. The unfilled histograms are for the jets
1790
+ clustered using the DR-AK algorithm, and the filled ones are for the jets using the AK
1791
+ clustering algorithm. Any untagged jet with pT > 5 GeV was considered to be a soft jet.
1792
+ Fig. 11(b) shows the normalized distribution for the finally acquired radii of different jets
1793
+ for category C22 events. The desirable feature of the reconstructed W jets being narrower
1794
+ than the reconstructed top jets is clearly apparent in the figure. Here, again, the soft jets
1795
+ are growing to larger radius are expected. However, as discussed in the previous section,
1796
+ they can be removed from an analysis by pT or jet mass cuts.
1797
+ The values of A [Eq. (3.1)] for the two different algorithms, viz. DR-AK and AK are
1798
+ also quoted in Fig. 11(a). These values (1.62% for AK and 5.47% for DR-AK), clearly,
1799
+ indicate that the dynamic radius jet algorithm is working better while probing the correct
1800
+ mass windows for the particles. The shift of the mass distribution towards larger values is
1801
+ indicative of capturing little extra than required. As discussed previously, this can be recti-
1802
+ fied by the techniques of grooming [77–81], trimming [84], or pruning [85, 86]. Furthermore,
1803
+ going beyond just the jet mass to tag the topor W jets would further help in extracting
1804
+ signals.
1805
+ The variation of A as a function of initial radius R0 is shown in Fig. 12 for six categories,
1806
+ namely C22, C21, C20, C13, C12, and C11. These categories have at least one top jet
1807
+ identified within 0.5 distance from the MG5 parton-level top quark. The solid blue lines
1808
+ – 23 –
1809
+
1810
+ 0.2
1811
+ 0.3
1812
+ 0.4
1813
+ 0.5
1814
+ 0.6
1815
+ 0.7
1816
+ 0.8
1817
+ R0
1818
+ 0
1819
+ 5
1820
+ 10
1821
+ 15
1822
+ 20
1823
+ Acceptance [%]
1824
+ C22
1825
+ DR-AK
1826
+ AK
1827
+ 0.2
1828
+ 0.3
1829
+ 0.4
1830
+ 0.5
1831
+ 0.6
1832
+ 0.7
1833
+ 0.8
1834
+ R0
1835
+ 0
1836
+ 5
1837
+ 10
1838
+ 15
1839
+ 20
1840
+ Acceptance [%]
1841
+ C21
1842
+ DR-AK
1843
+ AK
1844
+ 0.2
1845
+ 0.3
1846
+ 0.4
1847
+ 0.5
1848
+ 0.6
1849
+ 0.7
1850
+ 0.8
1851
+ R0
1852
+ 0
1853
+ 5
1854
+ 10
1855
+ 15
1856
+ 20
1857
+ Acceptance [%]
1858
+ C20
1859
+ DR-AK
1860
+ AK
1861
+ 0.2
1862
+ 0.3
1863
+ 0.4
1864
+ 0.5
1865
+ 0.6
1866
+ 0.7
1867
+ 0.8
1868
+ R0
1869
+ 0
1870
+ 5
1871
+ 10
1872
+ 15
1873
+ 20
1874
+ Acceptance [%]
1875
+ C13
1876
+ DR-AK
1877
+ AK
1878
+ 0.2
1879
+ 0.3
1880
+ 0.4
1881
+ 0.5
1882
+ 0.6
1883
+ 0.7
1884
+ 0.8
1885
+ R0
1886
+ 0
1887
+ 5
1888
+ 10
1889
+ 15
1890
+ 20
1891
+ Acceptance [%]
1892
+ C12
1893
+ DR-AK
1894
+ AK
1895
+ 0.2
1896
+ 0.3
1897
+ 0.4
1898
+ 0.5
1899
+ 0.6
1900
+ 0.7
1901
+ 0.8
1902
+ R0
1903
+ 0
1904
+ 5
1905
+ 10
1906
+ 15
1907
+ 20
1908
+ Acceptance [%]
1909
+ C11
1910
+ DR-AK
1911
+ AK
1912
+ Figure 12: The variation of A [Eq. (3.1)] as a function of initial radius R0 for six categories,
1913
+ namely C22, C21, C20, C13, C12, and C11.
1914
+ The solid blue lines are for the DR-AK
1915
+ algorithm, and the dashed lines are for the AK algorithm. The jets are clustered with
1916
+ R0 = 0.5.
1917
+ represent the efficiencies for the DR-AK algorithm, and the dashed lines are representative
1918
+ of the AK algorithm. For the case of dynamic radius, the quintessential feature is the
1919
+ initial increment in the acceptance efficiencies A up to R0 = 0.5, and, beyond this value,
1920
+ the efficiencies decrease. The reason for this is an unnecessary accumulation of hadrons
1921
+ and making the jets bigger than their optimal size. However, for the AK algorithm, the
1922
+ efficiencies keep on increasing until R0 = 0.7, which is kind of the optimal radius for this
1923
+ scenario.
1924
+ The most important point to note is that up to R0=0.5, the efficiencies for
1925
+ the DR-AK algorithm are higher than those for the AK algorithm. This feature, again,
1926
+ establishes the utility of using dynamic radius algorithms over fixed radius ones.
1927
+ In the end, we look at the bar plot of the acceptance efficiencies A for all the categories
1928
+ in Fig. 13. The blue and green bars are for DR-AK and AK algorithms, respectively. The
1929
+ initial radius R0 is taken to be 0.5. The numbers under the curly braces below the x-axis
1930
+ represent the values of A for the categories which capture 2 tops, 1 top, 0 top, and none of
1931
+ the top or W jets. The important observation in this regard is that the categories containing
1932
+ 2 top and 1 top jets have better efficiencies for the dynamic radius algorithm than the fixed
1933
+ radius one. This means that the events, where the AK algorithm could not capture the
1934
+ whole of the top constituents, the DR-AK algorithm could capture the full tops. Thus the
1935
+ credence of our proposed algorithm is established in a BSM context as well.
1936
+ – 24 –
1937
+
1938
+ C22 C21 C20 C13 C12 C11 C10 C04 C03 C02 C01 C00
1939
+ Categories
1940
+ 0
1941
+ 5
1942
+ 10
1943
+ 15
1944
+ 20
1945
+ 25
1946
+ 30
1947
+ Acceptance [%]
1948
+ R0 = 0.5
1949
+ DR-AK:
1950
+ AK:
1951
+
1952
+ ��
1953
+
1954
+ 9.86%
1955
+ 4.38%
1956
+
1957
+ ��
1958
+
1959
+ 35.33%
1960
+ 27.31%
1961
+
1962
+ ��
1963
+
1964
+ 34.02%
1965
+ 53.75%
1966
+ ����
1967
+ 20.79%
1968
+ 14.56%
1969
+ DR-AK
1970
+ AK
1971
+ Figure 13: Bar plot of A for different categories for jet algorithms with R0 = 0.5. The
1972
+ blue, and green bars are for DR-AK and AK algorithm respectively. From left to right,
1973
+ The numbers under the braces represent the values of A for the categories which capture
1974
+ 2 top, 1 top, 0 top, and none of the top or W jets.
1975
+ 5
1976
+ Summary and Outlook
1977
+ We go beyond the most popular jet clustering algorithms, where the formation of jets
1978
+ is performed using a fixed radius parameter.
1979
+ These algorithms return fixed-sized jets
1980
+ corresponding to the input radius parameter. In this work, an attempt is made to make
1981
+ the radius of each jet variable depending on the kinematics and hadronic activity in the
1982
+ neighbourhood of an evolving jet. The proposed method is based on the standard kt-type
1983
+ sequential recombination jet clustering algorithms with the incorporation of the dynamic
1984
+ nature of the radius parameter.
1985
+ Starting from a reasonable radius parameter, during the process of formation of a jet,
1986
+ the radius of each evolving jet is allowed to grow based on fuzziness inside it. For this
1987
+ work, the measure of the fuzziness of each evolving jet is chosen to be the ‘pT -weighted’
1988
+ standard deviation of the inter-particle distances (in the η-φ plane) of the particles inside
1989
+ the evolving jet.
1990
+ After describing the proposed method, we have presented two different SM processes,
1991
+ viz. pp → tj and pp → Wj +Zj, to demonstrate some applicabilities of the dynamic radius
1992
+ jet clustering algorithm. In these two processes, differently-sized jets are expected in a sin-
1993
+ gle event. In the two SM process examples, we observe that the jets are being formed with
1994
+ radii varying in size on a jet-by-jet basis. In terms of the acceptance efficiency [Eq. (3.1)],
1995
+ we show that the performance of the dynamic radius algorithm is better compared to their
1996
+ – 25 –
1997
+
1998
+ fixed radius counterparts. We take up a scenario beyond the Standard Model for further
1999
+ illustration, where a vectorlike SU(2)L singlet charge −1/3 quark b′ is added. We study
2000
+ jet clustering in pp → b′¯b′ followed by each b′ decaying into tW. Once more, our proposed
2001
+ method turns out to be effective in the reconstruction of the final state particles.
2002
+ In the examples given above, the dynamicity has been incorporated in the radius
2003
+ parameter of the standard kt-type sequential recombination algorithm. The central idea
2004
+ is the usage of fuzziness of an evolving jet to appropriately increase its radius starting
2005
+ from a starting radius R0. Although examples with only one measure of fuzziness have
2006
+ been shown in this work, one may consider other appropriate measures. depending upon
2007
+ the underlying physics process or the final goal of the analysis. Therefore, the idea of the
2008
+ dynamic radius jet algorithm should not be restricted only to this particular measure. The
2009
+ applicability of these possibilities will be presented in a separate work. In a nutshell, the
2010
+ idea of dynamic radius jet clustering algorithm on a jet-by-jet basis is useful in collider
2011
+ studies and will be beneficial in searches driven by processes in SM as well as BSM.
2012
+ Acknowledgments
2013
+ The authors thank Jayita Lahiri for useful discussions during the initial phase of the work.
2014
+ References
2015
+ [1] J.M. Campbell, J.W. Huston and W.J. Stirling, Hard Interactions of Quarks and Gluons: A
2016
+ Primer for LHC Physics, Rept. Prog. Phys. 70 (2007) 89 [hep-ph/0611148].
2017
+ [2] R.K. Ellis, W.J. Stirling and B.R. Webber, QCD and collider physics, vol. 8, Cambridge
2018
+ University Press (2, 2011), 10.1017/CBO9780511628788.
2019
+ [3] G.F. Sterman and S. Weinberg, Jets from Quantum Chromodynamics, Phys. Rev. Lett. 39
2020
+ (1977) 1436.
2021
+ [4] J.E. Huth et al., Toward a standardization of jet definitions, in 1990 DPF Summer Study
2022
+ on High-energy Physics: Research Directions for the Decade (Snowmass 90), pp. 0134–136,
2023
+ 12, 1990.
2024
+ [5] G.C. Blazey et al., Run II jet physics, in Physics at Run II: QCD and Weak Boson Physics
2025
+ Workshop: Final General Meeting, pp. 47–77, 5, 2000 [hep-ex/0005012].
2026
+ [6] S.D. Ellis, J. Huston, K. Hatakeyama, P. Loch and M. Tonnesmann, Jets in hadron-hadron
2027
+ collisions, Prog. Part. Nucl. Phys. 60 (2008) 484 [0712.2447].
2028
+ [7] G.P. Salam, Towards Jetography, Eur. Phys. J. C 67 (2010) 637 [0906.1833].
2029
+ [8] S. Catani, Y.L. Dokshitzer, M. Olsson, G. Turnock and B.R. Webber, New clustering
2030
+ algorithm for multi - jet cross-sections in e+ e- annihilation, Phys. Lett. B 269 (1991) 432.
2031
+ [9] S. Catani, Y.L. Dokshitzer, M.H. Seymour and B.R. Webber, Longitudinally invariant Kt
2032
+ clustering algorithms for hadron hadron collisions, Nucl. Phys. B 406 (1993) 187.
2033
+ [10] S.D. Ellis and D.E. Soper, Successive combination jet algorithm for hadron collisions, Phys.
2034
+ Rev. D 48 (1993) 3160 [hep-ph/9305266].
2035
+ [11] Y.L. Dokshitzer, G.D. Leder, S. Moretti and B.R. Webber, Better jet clustering algorithms,
2036
+ JHEP 08 (1997) 001 [hep-ph/9707323].
2037
+ – 26 –
2038
+
2039
+ [12] M. Wobisch and T. Wengler, Hadronization corrections to jet cross-sections in deep
2040
+ inelastic scattering, in Workshop on Monte Carlo Generators for HERA Physics (Plenary
2041
+ Starting Meeting), pp. 270–279, 4, 1998 [hep-ph/9907280].
2042
+ [13] M. Cacciari, G.P. Salam and G. Soyez, The anti-kt jet clustering algorithm, JHEP 04
2043
+ (2008) 063 [0802.1189].
2044
+ [14] A.J. Larkoski, J. Thaler and W.J. Waalewijn, Gaining (Mutual) Information about
2045
+ Quark/Gluon Discrimination, JHEP 11 (2014) 129 [1408.3122].
2046
+ [15] P. Gras, S. H¨oche, D. Kar, A. Larkoski, L. L¨onnblad, S. Pl¨atzer et al., Systematics of
2047
+ quark/gluon tagging, JHEP 07 (2017) 091 [1704.03878].
2048
+ [16] J. Gallicchio and M.D. Schwartz, Quark and Gluon Tagging at the LHC, Phys. Rev. Lett.
2049
+ 107 (2011) 172001 [1106.3076].
2050
+ [17] J. Gallicchio and M.D. Schwartz, Quark and Gluon Jet Substructure, JHEP 04 (2013) 090
2051
+ [1211.7038].
2052
+ [18] J.R. Andersen et al., Les Houches 2015: Physics at TeV Colliders Standard Model Working
2053
+ Group Report, in 9th Les Houches Workshop on Physics at TeV Colliders, 5, 2016
2054
+ [1605.04692].
2055
+ [19] B. Bhattacherjee, S. Mukhopadhyay, M.M. Nojiri, Y. Sakaki and B.R. Webber, Quark-gluon
2056
+ discrimination in the search for gluino pair production at the LHC, JHEP 01 (2017) 044
2057
+ [1609.08781].
2058
+ [20] P.T. Komiske, E.M. Metodiev and J. Thaler, An operational definition of quark and gluon
2059
+ jets, JHEP 11 (2018) 059 [1809.01140].
2060
+ [21] S. Bright-Thonney and B. Nachman, Investigating the Topology Dependence of Quark and
2061
+ Gluon Jets, JHEP 03 (2019) 098 [1810.05653].
2062
+ [22] A.J. Larkoski and E.M. Metodiev, A Theory of Quark vs. Gluon Discrimination, JHEP 10
2063
+ (2019) 014 [1906.01639].
2064
+ [23] A.K. Nayak, S.K. Rai and T. Samui, Improving Heavy Dijet Resonance Searches Using Jet
2065
+ Substructure at the LHC, Eur. Phys. J. C 81 (2021) 130 [1912.03511].
2066
+ [24] D. Reichelt, S. Caletti, O. Fedkevych, S. Marzani, S. Schumann and G. Soyez,
2067
+ Phenomenology of jet angularities at the LHC, JHEP 03 (2022) 131 [2112.09545].
2068
+ [25] I.W. Stewart and X. Yao, Pure quark and gluon observables in collinear drop, JHEP 09
2069
+ (2022) 120 [2203.14980].
2070
+ [26] S. Bright-Thonney, I. Moult, B. Nachman and S. Prestel, Systematic quark/gluon
2071
+ identification with ratios of likelihoods, JHEP 12 (2022) 021 [2207.12411].
2072
+ [27] S.D. Ellis, A. Hornig, T.S. Roy, D. Krohn and M.D. Schwartz, Qjets: A Non-Deterministic
2073
+ Approach to Tree-Based Jet Substructure, Phys. Rev. Lett. 108 (2012) 182003 [1201.1914].
2074
+ [28] S.D. Ellis, A. Hornig, D. Krohn and T.S. Roy, On Statistical Aspects of Qjets, JHEP 01
2075
+ (2015) 022 [1409.6785].
2076
+ [29] Les Houches 2017: Physics at TeV Colliders Standard Model Working Group Report, 3,
2077
+ 2018.
2078
+ [30] S. Marzani, G. Soyez and M. Spannowsky, Looking inside jets: an introduction to jet
2079
+ substructure and boosted-object phenomenology, vol. 958, Springer (2019),
2080
+ 10.1007/978-3-030-15709-8, [1901.10342].
2081
+ – 27 –
2082
+
2083
+ [31] A.J. Larkoski, G.P. Salam and J. Thaler, Energy Correlation Functions for Jet
2084
+ Substructure, JHEP 06 (2013) 108 [1305.0007].
2085
+ [32] A.J. Larkoski, I. Moult and D. Neill, Analytic Boosted Boson Discrimination, JHEP 05
2086
+ (2016) 117 [1507.03018].
2087
+ [33] A.J. Larkoski, I. Moult and D. Neill, Power Counting to Better Jet Observables, JHEP 12
2088
+ (2014) 009 [1409.6298].
2089
+ [34] A.J. Larkoski, I. Moult and D. Neill, Building a Better Boosted Top Tagger, Phys. Rev. D
2090
+ 91 (2015) 034035 [1411.0665].
2091
+ [35] D. Adams et al., Towards an Understanding of the Correlations in Jet Substructure, Eur.
2092
+ Phys. J. C 75 (2015) 409 [1504.00679].
2093
+ [36] I. Moult, L. Necib and J. Thaler, New Angles on Energy Correlation Functions, JHEP 12
2094
+ (2016) 153 [1609.07483].
2095
+ [37] R.S. Chivukula, K.A. Mohan, D. Sengupta and E.H. Simmons, Characterizing boosted dijet
2096
+ resonances with energy correlation functions, JHEP 03 (2018) 133 [1710.04661].
2097
+ [38] J. Thaler and K. Van Tilburg, Identifying Boosted Objects with N-subjettiness, JHEP 03
2098
+ (2011) 015 [1011.2268].
2099
+ [39] I.W. Stewart, F.J. Tackmann, J. Thaler, C.K. Vermilion and T.F. Wilkason, XCone:
2100
+ N-jettiness as an Exclusive Cone Jet Algorithm, JHEP 11 (2015) 072 [1508.01516].
2101
+ [40] A. Dey and T. Samui, Jet Substructure and Multivariate Analysis Aid in Polarization Study
2102
+ of Boosted, Hadronic W Fatjet at the LHC, 2110.02773.
2103
+ [41] J. Thaler and K. Van Tilburg, Maximizing Boosted Top Identification by Minimizing
2104
+ N-subjettiness, JHEP 02 (2012) 093 [1108.2701].
2105
+ [42] Z. Han, Tracking the Identities of Boosted Particles, Phys. Rev. D 86 (2012) 014026
2106
+ [1112.3378].
2107
+ [43] T. Plehn and M. Spannowsky, Top Tagging, J. Phys. G 39 (2012) 083001 [1112.4441].
2108
+ [44] G. Kasieczka, T. Plehn, T. Schell, T. Strebler and G.P. Salam, Resonance Searches with an
2109
+ Updated Top Tagger, JHEP 06 (2015) 203 [1503.05921].
2110
+ [45] M. Dasgupta, L. Schunk and G. Soyez, Jet shapes for boosted jet two-prong decays from
2111
+ first-principles, JHEP 04 (2016) 166 [1512.00516].
2112
+ [46] G.P. Salam, L. Schunk and G. Soyez, Dichroic subjettiness ratios to distinguish colour flows
2113
+ in boosted boson tagging, JHEP 03 (2017) 022 [1612.03917].
2114
+ [47] Z. Han, M. Son and B. Tweedie, Top-Tagging at the Energy Frontier, Phys. Rev. D 97
2115
+ (2018) 036023 [1707.06741].
2116
+ [48] J.A. Aguilar-Saavedra, J.H. Collins and R.K. Mishra, A generic anti-QCD jet tagger, JHEP
2117
+ 11 (2017) 163 [1709.01087].
2118
+ [49] A. Das, P. Konar and A. Thalapillil, Jet substructure shedding light on heavy Majorana
2119
+ neutrinos at the LHC, JHEP 02 (2018) 083 [1709.09712].
2120
+ [50] L. Moore, K. Nordstr¨om, S. Varma and M. Fairbairn, Reports of My Demise Are Greatly
2121
+ Exaggerated: N-subjettiness Taggers Take On Jet Images, SciPost Phys. 7 (2019) 036
2122
+ [1807.04769].
2123
+ – 28 –
2124
+
2125
+ [51] D. Napoletano and G. Soyez, Computing N-subjettiness for boosted jets, JHEP 12 (2018)
2126
+ 031 [1809.04602].
2127
+ [52] K. Agashe, J.H. Collins, P. Du, S. Hong, D. Kim and R.K. Mishra, Detecting a Boosted
2128
+ Diboson Resonance, JHEP 11 (2018) 027 [1809.07334].
2129
+ [53] L. Bradshaw, R.K. Mishra, A. Mitridate and B. Ostdiek, Mass Agnostic Jet Taggers,
2130
+ SciPost Phys. 8 (2020) 011 [1908.08959].
2131
+ [54] J.A. Aguilar-Saavedra and B. Zald´ıvar, Jet tagging made easy, Eur. Phys. J. C 80 (2020)
2132
+ 530 [2002.12320].
2133
+ [55] Y. Mehtar-Tani, A. Soto-Ontoso and K. Tywoniuk, Tagging boosted hadronic objects with
2134
+ dynamical grooming, Phys. Rev. D 102 (2020) 114013 [2005.07584].
2135
+ [56] A. Bhardwaj, J. Dutta, P. Konar, B. Mukhopadhyaya and S.K. Rai, Boosted jet techniques
2136
+ for a supersymmetric scenario with gravitino LSP, JHEP 10 (2020) 083 [2007.00351].
2137
+ [57] S. De, V. Rentala and W. Shepherd, Measuring the polarization of boosted, hadronic W
2138
+ bosons with jet substructure observables, 2008.04318.
2139
+ [58] S. Bhattacharya, M. Guchait and A.H. Vijay, Boosted top quark tagging and polarization
2140
+ measurement using machine learning, Phys. Rev. D 105 (2022) 042005 [2010.11778].
2141
+ [59] D. Choudhury, K. Deka and N. Kumar, Looking for a vectorlike B quark at the LHC using
2142
+ jet substructure, Phys. Rev. D 104 (2021) 035004 [2103.10655].
2143
+ [60] M. Dasgupta and J. Helliwell, Investigating top tagging with Ym-Splitter and N-subjettiness,
2144
+ JHEP 10 (2021) 092 [2108.09317].
2145
+ [61] Y. Lu, A. Romero, M.J. Fenton, D. Whiteson and P. Baldi, Resolving extreme jet
2146
+ substructure, JHEP 08 (2022) 046 [2202.00723].
2147
+ [62] A. Bhardwaj, K. Bhide, T. Mandal, S. Mitra and C. Neeraj, Discovery prospects of a
2148
+ vectorlike top partner decaying to a singlet boson, Phys. Rev. D 106 (2022) 075024
2149
+ [2204.09005].
2150
+ [63] A. Dey, R. Rahaman and S.K. Rai, Fatjet signatures of heavy neutrinos and heavy leptons
2151
+ in a left-right model with universal seesaw at the HL-LHC, 2207.06857.
2152
+ [64] B. Bhattacherjee, C. Bose, A. Chakraborty and R. Sengupta, Boosted top tagging and its
2153
+ interpretation using Shapley values, 2212.11606.
2154
+ [65] P.K. Das, P. Konar, S. Kundu and S. Show, Jet substructure probe to unfold singlet-doublet
2155
+ dark matter in the presence of non-standard cosmology, 2301.02514.
2156
+ [66] CMS collaboration, Measurement of jet substructure observables in tt events from
2157
+ proton-proton collisions at √s = 13TeV, Phys. Rev. D 98 (2018) 092014 [1808.07340].
2158
+ [67] ATLAS collaboration, Measurement of jet-substructure observables in top quark, W boson
2159
+ and light jet production in proton-proton collisions at √s = 13 TeV with the ATLAS
2160
+ detector, JHEP 08 (2019) 033 [1903.02942].
2161
+ [68] CMS collaboration, Study of quark and gluon jet substructure in Z+jet and dijet events
2162
+ from pp collisions, JHEP 01 (2022) 188 [2109.03340].
2163
+ [69] D. Krohn, J. Thaler and L.-T. Wang, Jets with Variable R, JHEP 06 (2009) 059
2164
+ [0903.0392].
2165
+ – 29 –
2166
+
2167
+ [70] L. Mackey, B. Nachman, A. Schwartzman and C. Stansbury, Fuzzy Jets, JHEP 06 (2016)
2168
+ 010 [1509.02216].
2169
+ [71] M. Cacciari, G.P. Salam and G. Soyez, FastJet User Manual, Eur. Phys. J. C 72 (2012)
2170
+ 1896 [1111.6097].
2171
+ [72] M. Cacciari and G.P. Salam, Dispelling the N 3 myth for the kt jet-finder, Phys. Lett. B 641
2172
+ (2006) 57 [hep-ph/0512210].
2173
+ [73] J. Alwall, R. Frederix, S. Frixione, V. Hirschi, F. Maltoni, O. Mattelaer et al., The
2174
+ automated computation of tree-level and next-to-leading order differential cross sections,
2175
+ and their matching to parton shower simulations, JHEP 07 (2014) 079 [1405.0301].
2176
+ [74] T. Sj¨ostrand, S. Mrenna and P.Z. Skands, PYTHIA 6.4 Physics and Manual, JHEP 05
2177
+ (2006) 026 [hep-ph/0603175].
2178
+ [75] T. Sj¨ostrand, S. Ask, J.R. Christiansen, R. Corke, N. Desai, P. Ilten et al., An introduction
2179
+ to PYTHIA 8.2, Comput. Phys. Commun. 191 (2015) 159 [1410.3012].
2180
+ [76] P. Skands, S. Carrazza and J. Rojo, Tuning PYTHIA 8.1: the Monash 2013 Tune, Eur.
2181
+ Phys. J. C 74 (2014) 3024 [1404.5630].
2182
+ [77] J.M. Butterworth, A.R. Davison, M. Rubin and G.P. Salam, Jet substructure as a new
2183
+ Higgs search channel at the LHC, Phys. Rev. Lett. 100 (2008) 242001 [0802.2470].
2184
+ [78] M. Dasgupta, A. Fregoso, S. Marzani and G.P. Salam, Towards an understanding of jet
2185
+ substructure, JHEP 09 (2013) 029 [1307.0007].
2186
+ [79] A.J. Larkoski, S. Marzani, G. Soyez and J. Thaler, Soft Drop, JHEP 05 (2014) 146
2187
+ [1402.2657].
2188
+ [80] S. Marzani, L. Schunk and G. Soyez, A study of jet mass distributions with grooming, JHEP
2189
+ 07 (2017) 132 [1704.02210].
2190
+ [81] S. Marzani, L. Schunk and G. Soyez, The jet mass distribution after Soft Drop, Eur. Phys.
2191
+ J. C 78 (2018) 96 [1712.05105].
2192
+ [82] F.A. Dreyer, L. Necib, G. Soyez and J. Thaler, Recursive Soft Drop, JHEP 06 (2018) 093
2193
+ [1804.03657].
2194
+ [83] Y. Mehtar-Tani, A. Soto-Ontoso and K. Tywoniuk, Dynamical grooming of QCD jets, Phys.
2195
+ Rev. D 101 (2020) 034004 [1911.00375].
2196
+ [84] D. Krohn, J. Thaler and L.-T. Wang, Jet Trimming, JHEP 02 (2010) 084 [0912.1342].
2197
+ [85] S.D. Ellis, C.K. Vermilion and J.R. Walsh, Techniques for improved heavy particle searches
2198
+ with jet substructure, Phys. Rev. D 80 (2009) 051501 [0903.5081].
2199
+ [86] S.D. Ellis, C.K. Vermilion and J.R. Walsh, Recombination Algorithms and Jet Substructure:
2200
+ Pruning as a Tool for Heavy Particle Searches, Phys. Rev. D 81 (2010) 094023 [0912.0033].
2201
+ [87] V.D. Barger, N. Deshpande, R.J.N. Phillips and K. Whisnant, Extra Fermions in E6
2202
+ Superstring Theories, Phys. Rev. D 33 (1986) 1912.
2203
+ [88] T.G. Rizzo, Phenomenology of Exotic Particles in E(6) Theories, Phys. Rev. D 34 (1986)
2204
+ 1438.
2205
+ [89] B. Mukhopadhyaya and S. Nandi, Evading the top mass bound at the Tevatron: New signals
2206
+ for the top, Phys. Rev. Lett. 66 (1991) 285.
2207
+ – 30 –
2208
+
2209
+ [90] B. Mukhopadhyaya and S. Nandi, Collider implications of singlet fermions, Phys. Rev. D
2210
+ 46 (1992) 5098.
2211
+ [91] R.S. Chivukula, B.A. Dobrescu, H. Georgi and C.T. Hill, Top Quark Seesaw Theory of
2212
+ Electroweak Symmetry Breaking, Phys. Rev. D 59 (1999) 075003 [hep-ph/9809470].
2213
+ [92] M. Gillioz, R. Gr¨ober, A. Kapuvari and M. M¨uhlleitner, Vector-like Bottom Quarks in
2214
+ Composite Higgs Models, JHEP 03 (2014) 037 [1311.4453].
2215
+ [93] J. Maalampi and M. Roos, Flavor Mixing in the Presence of a Fourth Down Quark, Phys.
2216
+ Lett. B 188 (1987) 487.
2217
+ [94] A. Raychaudhuri, Is the b Quark Mixed with a Vector Like Quark, Phys. Rev. D 40 (1989)
2218
+ 833.
2219
+ [95] V.D. Barger, M.S. Berger and R.J.N. Phillips, Quark singlets: Implications and constraints,
2220
+ Phys. Rev. D 52 (1995) 1663 [hep-ph/9503204].
2221
+ [96] T.C. Andre and J.L. Rosner, Exotic Q = -1/3 quark signatures at high-energy hadron
2222
+ colliders, Phys. Rev. D 69 (2004) 035009 [hep-ph/0309254].
2223
+ [97] S. Gopalakrishna, T. Mandal, S. Mitra and R. Tibrewala, LHC Signatures of a Vector-like
2224
+ b’, Phys. Rev. D 84 (2011) 055001 [1107.4306].
2225
+ [98] J.A. Aguilar-Saavedra, R. Benbrik, S. Heinemeyer and M. P´erez-Victoria, Handbook of
2226
+ vectorlike quarks: Mixing and single production, Phys. Rev. D 88 (2013) 094010
2227
+ [1306.0572].
2228
+ [99] A. Girdhar, B. Mukhopadhyaya and M. Patra, Distinguishing Signatures of top-and
2229
+ bottom-type heavy vectorlike quarks at the LHC, Phys. Rev. D 91 (2015) 055015
2230
+ [1404.3374].
2231
+ [100] S.A.R. Ellis, R.M. Godbole, S. Gopalakrishna and J.D. Wells, Survey of vector-like fermion
2232
+ extensions of the Standard Model and their phenomenological implications, JHEP 09 (2014)
2233
+ 130 [1404.4398].
2234
+ [101] K. Das, T. Li, S. Nandi and S.K. Rai, New signals for vector-like down-type quark in U(1)
2235
+ of E6, Eur. Phys. J. C 78 (2018) 35 [1708.00328].
2236
+ [102] CMS collaboration, A search for bottom-type, vector-like quark pair production in a fully
2237
+ hadronic final state in proton-proton collisions at √s = 13 TeV, Phys. Rev. D 102 (2020)
2238
+ 112004 [2008.09835].
2239
+ [103] CMS collaboration, Search for pair production of vector-like quarks in leptonic final states
2240
+ in proton-proton collisions at √s = 13 TeV, 2209.07327.
2241
+ [104] ATLAS collaboration, Search for pair-production of vector-like quarks in pp collision events
2242
+ at √s = 13 TeV with at least one leptonically decaying Z boson and a third-generation
2243
+ quark with the ATLAS detector, 2210.15413.
2244
+ [105] ATLAS collaboration, Search for pair-produced vector-like top and bottom partners in
2245
+ events with large missing transverse momentum in pp collisions with the ATLAS detector,
2246
+ 2212.05263.
2247
+ [106] B. Mukhopadhyaya, T. Samui and R.K. Singh, in preparation.
2248
+ [107] F. Staub, SARAH, 0806.0538.
2249
+ [108] F. Staub, SARAH 4 : A tool for (not only SUSY) model builders, Comput. Phys. Commun.
2250
+ 185 (2014) 1773 [1309.7223].
2251
+ – 31 –
2252
+
2253
+ [109] F. Staub, Exploring new models in all detail with SARAH, Adv. High Energy Phys. 2015
2254
+ (2015) 840780 [1503.04200].
2255
+ [110] C. Degrande, C. Duhr, B. Fuks, D. Grellscheid, O. Mattelaer and T. Reiter, UFO - The
2256
+ Universal FeynRules Output, Comput. Phys. Commun. 183 (2012) 1201 [1108.2040].
2257
+ [111] W. Porod, SPheno, a program for calculating supersymmetric spectra, SUSY particle decays
2258
+ and SUSY particle production at e+ e- colliders, Comput. Phys. Commun. 153 (2003) 275
2259
+ [hep-ph/0301101].
2260
+ [112] W. Porod and F. Staub, SPheno 3.1: Extensions including flavour, CP-phases and models
2261
+ beyond the MSSM, Comput. Phys. Commun. 183 (2012) 2458 [1104.1573].
2262
+ – 32 –
2263
+
9dFPT4oBgHgl3EQfYjT_/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
9tE4T4oBgHgl3EQfDQub/content/tmp_files/2301.04868v1.pdf.txt ADDED
@@ -0,0 +1,2267 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ MNRAS 000, 1–15 (20XX)
2
+ Preprint 13 January 2023
3
+ Compiled using MNRAS LATEX style file v3.0
4
+ Investigating Dynamical Properties of Globular Clusters through a Family
5
+ of Lowered Isothermal Models
6
+ Chia-Hsuan Cheng1 and Ing-Guey Jiang1,2
7
+ 1Department of Physics, National Tsing-Hua University, Hsinchu, Taiwan
8
+ 2Institute of Astronomy, National Tsing-Hua University, Hsinchu, Taiwan
9
+ Accepted XXX. Received YYY; in original form ZZZ
10
+ ABSTRACT
11
+ To investigate the dynamical properties of globular clusters, the surface brightness and kinematic data were collected and fitted
12
+ to a family of lowered isothermal models called LIMEPY models. For 18 studied globular clusters, the amounts of concentration,
13
+ truncation, and anisotropy were determined. In addition, the cluster mass, half-mass radius, distance, and mass-to-light ratio were
14
+ also obtained. In general, LIMEPY models could describe these clusters well. Among these 18 clusters, NGC 5139, NGC 6388, and
15
+ NGC 7078 were claimed to be candidates to host intermediate-mass black holes in literature. The models could not appropriately
16
+ fit the central proper-motion velocity dispersion of NGC 5139 and the slope of proper-motion velocity-dispersion profile of NGC
17
+ 6388. Thus, more dedicated models with intermediate-mass black holes or a group of stellar-mass black holes at cluster centers
18
+ may need to be considered. Considering NGC 7078, our model with some degree of anisotropy can fit the data. Finally, the
19
+ strong concentration-truncation anti-correlation and truncation-semimajor-axis correlation were revealed, which could be the
20
+ observational imprint of the dynamical evolution of globular clusters.
21
+ Key words: methods: numerical – stars: kinematics and dynamics – globular clusters: general – globular clusters: individual –
22
+ galaxies: star clusters: general
23
+ 1 INTRODUCTION
24
+ Globular clusters are one of the oldest objects in the universe (Van-
25
+ denberg et al. 1996). They extend spherically in several or tens of
26
+ parsecs with hundreds of thousands of stars (Harris 1996). The high
27
+ stellar densities make them the primary venue for hosting exotic
28
+ objects like millisecond pulsars (Manchester et al. 1991) and blue
29
+ stragglers (Bailyn 1995). Globular clusters have been proposed to
30
+ possibly also host intermediate-mass black holes (Ebisuzaki et al.
31
+ 2001). With higher density, the core of a globular cluster relaxes
32
+ faster than the halo and the relaxation time is short compared to
33
+ the age of the cluster (Oort & van Herk 1959). Thus, the center of
34
+ globular clusters is expected to be isothermal.
35
+ Having theoretical models describing globular clusters is help-
36
+ ful in obtaining the physical quantities. The isothermal sphere is a
37
+ model with isothermal cores, so it could be considered a suitable
38
+ simple model. However, this model extends to the infinite and has an
39
+ unrealistic infinite mass. This problem can be solved by introducing
40
+ some cutoffs. For example, energy truncation can limit the velocity,
41
+ so the stars with larger velocities escape from the cluster; this results
42
+ in a cluster model with finite mass and range. The truncation can
43
+ be regarded as the effect of the external tidal field on star clusters.
44
+ Different truncations lead to different models. For example, subtract-
45
+ ing a constant from the energy leads to the Woolley model (Woolley
46
+ 1954), and further subtraction from the distribution function gives
47
+ the King model (King 1966).
48
+ The velocity distributions of clusters in the above models are
49
+ isotropic. However, for realistic models, the possible anisotropy shall
50
+ be considered. The diffusion caused by stellar encounters facilitates
51
+ the entry of some stars into the cluster halo. These stars diffuse to
52
+ the halo along radial orbits and increase the radial anisotropy in the
53
+ halo (Spitzer & Shapiro 1972). The violent relaxation in the stage of
54
+ cluster formation can also contribute to some radial anisotropy in the
55
+ cluster halo (Lynden-Bell 1967). To include anisotropy in a model,
56
+ one can add the angular momentum into the distribution function.
57
+ The distribution function now depends on both the energy and the
58
+ angular momentum. For example, the Michie-King model (Michie
59
+ 1963) includes the angular momentum in an exponential term. This
60
+ model possesses the expected properties which contain an isothermal
61
+ core with some anisotropy at the outer parts.
62
+ A model with multi-mass components is another aspect of im-
63
+ provement. Da Costa & Freeman (1976) made the extension from the
64
+ King model by assuming that each component has the same form of
65
+ distribution function with different constants. Later, an anisotropic
66
+ multi-mass model was introduced by Gunn & Griffin (1979). Re-
67
+ cently, some extensions and unification of these isothermal models
68
+ have been developed. Considering the Woolley and the King model
69
+ as different schemes of energy truncation characterized by some in-
70
+ tegers, Gomez-Leyton & Velazquez (2014) established an extended
71
+ model which parametrized the truncation by a non-negative real
72
+ number. This was further generalized by Gieles & Zocchi (2015) to
73
+ include the radial anisotropy and multi-mass components in a fam-
74
+ ily of lowered isothermal models, which can cover more properties
75
+ of star clusters. They also provided a fast model solver written as a
76
+ Python code, LIMEPY, for this family of lowered isothermal models.
77
+ © 20XX The Authors
78
+ arXiv:2301.04868v1 [astro-ph.GA] 12 Jan 2023
79
+
80
+ 2
81
+ Cheng and Jiang
82
+ Thus, these models proposed by Gieles & Zocchi (2015) are called
83
+ LIMEPY models.
84
+ As presented by Zocchi et al. (2016), LIMEPY models could capture
85
+ the main properties of the globular clusters. Moreover, Zocchi et al.
86
+ (2017) applied LIMEPY models in the study of NGC 5139 and found
87
+ that part of the observed large central velocity dispersion could be
88
+ produced by anisotropic models. Thus, their results could provide
89
+ some constraints on the previously proposed central intermediate-
90
+ mass black hole in NGC 5139 (Noyola et al. 2010). This globular
91
+ cluster, also named 𝜔 Centauri, is the most complex one which has
92
+ many sub-populations (Sanna et al. 2020) and was heavily investi-
93
+ gated with many controversial results. On the other hand, the central
94
+ kinematics of NGC 6093 was studied by employing new integral-
95
+ field spectrograph data, and the existence of an intermediate-mass
96
+ black hole was supported (Göttgens et al. 2021). In addition, NGC
97
+ 6388 is also a candidate residence of the intermediate-mass black
98
+ hole (Lützgendorf et al. 2011).
99
+ Moreover, with Gaia data, Vasiliev & Baumgardt (2021) per-
100
+ formed a comprehensive study on the kinematic properties of many
101
+ Galactic globular clusters. The proper motions were measured and
102
+ the corresponding proper-motion dispersion profiles of 100 clusters
103
+ were obtained. Combining with HST and other literature data, Baum-
104
+ gardt & Vasiliev (2021) also accurately derived the distances to these
105
+ Galactic globular clusters.
106
+ Therefore, motivated by the development of LIMEPY models, the
107
+ controversial results of the central kinematics and intermediate-mass
108
+ black holes, and the availability of new data derived from the Gaia
109
+ mission, herein, we investigated the properties of 18 globular clusters
110
+ with the LIMEPY models. Including data from recent observations such
111
+ as the MUSE survey (Kamann et al. 2018) and Gaia mission (Vasiliev
112
+ & Baumgardt 2021), the physical parameters of these clusters were
113
+ obtained through the data-model fitting. Our results could lead to
114
+ updated and accurate descriptions of the dynamical states of these
115
+ clusters for the cases in which the data could be well fitted by the
116
+ LIMEPY models which can be isotropic or anisotropic. Our results
117
+ might also imply the possible existence of intermediate-mass black
118
+ holes for some globular clusters.
119
+ For the rest of this paper, in Section 2, we introduce the model’s
120
+ distribution function and essential properties. The observational data
121
+ are described in Section 3, and the parameter determination method
122
+ is shown in Section 4. The results and discussions are presented in
123
+ Section 5. In Section 6, some conclusions are made.
124
+ 2 THE MODEL
125
+ The LIMEPY models were employed as the standard model in this
126
+ study. As presented (Gieles & Zocchi 2015), there are single-mass
127
+ and multi-mass cases in LIMEPY models. Considering the single-mass
128
+ models, the distribution functions have the following form:
129
+ 𝑓 (𝐸, 𝐽) = 𝐴 exp
130
+ � −𝐽2
131
+ 2𝑟2a 𝑠2
132
+
133
+ 𝐸𝛾
134
+
135
+ 𝑔, 𝜙(𝑟t) − 𝐸
136
+ 𝑠2
137
+
138
+ ,
139
+ (1)
140
+ for 𝐸 ≤ 𝜙(𝑟t) and 𝑓 (𝐸, 𝐽) = 0 for 𝐸 > 𝜙(𝑟t). The function 𝐸𝛾(𝑔, 𝑥)
141
+ represents 𝑒𝑥 for 𝑔 = 0 and 𝑒𝑥𝛾(𝑔, 𝑥)/Γ(𝑔) for 𝑔 > 0, where 𝛾(𝑔, 𝑥)
142
+ is the lower incomplete gamma function and Γ(𝑔) stands for the
143
+ gamma function. This distribution function depends on the specific
144
+ energy 𝐸 and the specific angular momentum 𝐽. The function 𝜙 is the
145
+ gravitational potential and 𝑟t is the truncation radius. The parameter
146
+ 𝑔 is called the truncation parameter, and it regulates the energy
147
+ truncation of the model. The parameter 𝑟a is the anisotropic radius,
148
+ and it determines how anisotropic a system is. When 𝑟a grows, the
149
+ model is less anisotropic, and 𝑟a → ∞ corresponds to an isotropic
150
+ model. The constants 𝐴 and 𝑠 are used to set the physical scale of the
151
+ model. The density can be obtained by integrating the distribution
152
+ function 𝑓 (𝐸, 𝐽) over the velocity space:
153
+ 𝜌 =
154
+
155
+ 𝑓 (𝐸, 𝐽) d3𝑣.
156
+ (2)
157
+ Since 𝐸 = 𝑣2/2 + 𝜙(𝑟) and the distribution function is zero for 𝐸 >
158
+ 𝜙(𝑟t), it can be just integrated from 0 to 𝑣max = [2𝜙(𝑟t) − 2𝜙(𝑟)]1/2
159
+ at each 𝑟. This 𝑣max becomes zero when 𝑟 = 𝑟t and the density
160
+ vanishes for 𝑟 ≥ 𝑟t. Hence, the truncation radius 𝑟t represents the
161
+ distance where the density comes to zero.
162
+ The gravitational potential 𝜙 is subjected to the Poisson equation.
163
+ For spherical systems such as globular clusters, the equation results
164
+ in the following form:
165
+ d2𝜙
166
+ d𝑟2 + 2
167
+ 𝑟
168
+ d𝜙
169
+ d𝑟 = 4𝜋𝐺𝜌,
170
+ (3)
171
+ where 𝑟 is the radial coordinate and 𝐺 is the gravitational constant.
172
+ The relevant quantities were first turned into dimensionless ones for
173
+ solving the Poisson equation. The dimensionless potential is defined
174
+ as ˆ𝜙 = [𝜙(𝑟t) − 𝜙]/𝑠2. The dimensionless density and radius are
175
+ ˆ𝜌 = 𝜌/𝜌0 and ˆ𝑟 = 𝑟/𝑟0, where 𝜌0 and 𝑟0 satisfy 4𝜋𝐺𝑟2
176
+ 0𝜌0/𝑠2 = 9.
177
+ Then, the Poisson equation becomes
178
+ d2 ˆ𝜙
179
+ dˆ𝑟2 + 2
180
+ ˆ𝑟
181
+ d ˆ𝜙
182
+ dˆ𝑟 = −9 ˆ𝜌.
183
+ (4)
184
+ The equation is solved with the boundary conditions that, at ˆ𝑟 = 0,
185
+ d ˆ𝜙/dˆ𝑟 = 0 and ˆ𝜙 = 𝑊0, where 𝑊0 is a constant that specifies a
186
+ particular solution. Hence,𝑊0 is also a parameter of the LIMEPY model,
187
+ called the concentration parameter. It characterizes the concentration
188
+ of the model.
189
+ As previously mentioned, LIMEPY models provide an extended fam-
190
+ ily of isothermal models. Those famous models are included as sub-
191
+ families. For example, the Woolley model (Woolley 1954) can be
192
+ produced by setting 𝑔 = 0, 𝑟a → ∞. When 𝑔 = 1 and 𝑟a → ∞,
193
+ the King model (King 1966) is obtained. The Wilson model (Wilson
194
+ 1975), which is more extended, corresponds to 𝑔 = 2 and 𝑟a → ∞.
195
+ Models with 𝑊0 → ∞ or 𝑔 → ∞ become the isothermal spheres. In
196
+ addition, the polytrope can be represented as 𝑊0 → 0. It includes the
197
+ Plummer model (Plummer 1911) which corresponds to the model
198
+ with 𝑔 = 3.5. It has a finite mass but infinite extents. In general, the
199
+ model with appropriate 𝑊0 and 𝑟a can be finite in extent if 𝑔 < 3.5
200
+ and conversely infinite in extent with 𝑔 ≥ 3.5. In addition, Gieles &
201
+ Zocchi (2015) also showed that one kind of finite model is unsuitable
202
+ for star clusters. These systems have an upturn in the density far from
203
+ the center, so there is a large amount of mass in the halo. The ratio
204
+ of the virial radius and half-mass radius 𝑟v/𝑟h is a crucial parameter
205
+ for these models. They suggested that the models with 𝑟v/𝑟h ≥ 0.64
206
+ can adequately describe star clusters.
207
+ The LIMEPY models describe spherical systems with different con-
208
+ centrations, truncation, and radial anisotropy. In general, the model
209
+ is isotropic near the center but could be anisotropic in the middle
210
+ part of the system. The energy truncation limits the contribution of
211
+ anisotropy to radial orbits with 𝐸 ≈ 𝜙(𝑟t) and thus suppresses the
212
+ degree of radial anisotropy near the edge. The corresponding physi-
213
+ cal picture is that a cluster under the interaction of an external tidal
214
+ field has a preferential mass loss on stars with radial orbits. This re-
215
+ duces the amount of anisotropy in the outer region (Oh & Lin 1992;
216
+ Takahashi et al. 1997). Simulations of star clusters in the tidal field
217
+ confirmed this isotropic behavior near the edge (Tiongco et al. 2016).
218
+ Thus, the energy truncation acts as a role of the tidal field. In fact,
219
+ MNRAS 000, 1–15 (20XX)
220
+
221
+ Dynamical Properties of Globular Clusters
222
+ 3
223
+ the tidal field can also make the outer region profiles tangentially
224
+ anisotropic (Baumgardt & Makino 2003).
225
+ In addition to the anisotropic radius 𝑟a, there is a convenient
226
+ anisotropic parameter 𝜅 ≡ 2𝐾r/𝐾t, where 𝐾r is the total radial ki-
227
+ netic energy and 𝐾t is the total tangential kinetic energy. If 𝜅 > 1,
228
+ the system is radially anisotropic, and if 𝜅 < 1, the system is tangen-
229
+ tially anisotropic. When 𝜅 = 1, it is an isotropic system. Therefore, 𝜅
230
+ represents a simple and global measure of the anisotropy. We mainly
231
+ used 𝜅 to determine the amount of the anisotropy of clusters.
232
+ In Zocchi et al. (2016), the comparisons with N-body simulations
233
+ illustrated the variation of model parameters of a cluster during the
234
+ evolution. The cluster started with the Plummer model and the sim-
235
+ ulation snapshots at different time were fitted with LIMEPY models.
236
+ The concentration parameter tended to increase with time, which
237
+ was also suggested previously by King (1966). The truncation pa-
238
+ rameter 𝑔 decreased roughly from 2.5 to 0.5 during the evolution. It
239
+ corresponded to an increased truncation by the tidal field as a cluster
240
+ gradually filled the Roche volume. Thus, a cluster tends to become
241
+ more concentrated and truncated with time. In addition, the degree
242
+ of radial anisotropy increased due to radial diffusion but decreased
243
+ later during the core collapse.
244
+ 3 THE OBSERVATIONAL DATA
245
+ One of our primary goals is to provide updated results with a complete
246
+ inclusion of all available observational data for globular clusters. The
247
+ observational data of 𝑉-band surface brightness 𝜇 were taken from
248
+ Trager et al. (1995), which provided a catalog of surface brightness
249
+ profiles for over a hundred Galactic globular clusters. Some proce-
250
+ dures were needed before the data were ready for the fitting. There
251
+ was a correction related to extinction. The method is based on the
252
+ global mean curve discussed in Fitzpatrick (1999), which uses the
253
+ mean value for the ratio of the extinction ��𝑉 and the reddening
254
+ 𝐸(𝐵 − 𝑉) so that 𝐴𝑉 = 3.1𝐸(𝐵 − 𝑉). We took the reddening in
255
+ the catalog of Harris (1996) (2010 version) and then computed the
256
+ corrected surface brightness by 𝜇𝑖 = 𝜇𝑖,0 − 𝐴𝑉 , where 𝜇𝑖,0 denotes
257
+ the data before the correction. The data with 𝑤𝑖 < 0.15 were not
258
+ adopted according to McLaughlin & van der Marel (2005), where
259
+ 𝑤𝑖 is the weight of each data given in Trager et al. (1995).
260
+ Because the data number was large, which might make the surface
261
+ brightness dominate the fitting, we sliced the radial range with equal
262
+ logarithmic width and averaged the surface brightness and the weight
263
+ in each bin. The bin number was 55 which equaled the largest data
264
+ number of the velocity dispersion. To compute the uncertainty for
265
+ each data, we followed the method in McLaughlin & van der Marel
266
+ (2005). The uncertainty of the data was obtained by 𝜖𝜇,𝑖 = 𝜖𝜇,b/𝑤𝑖,
267
+ where 𝜖𝜇,b is the base error bar for each cluster.
268
+ For line-of-sight velocity dispersion, we used the profiles derived
269
+ from the collected literature (Baumgardt 2017), the data from un-
270
+ published spectra of stars in the ESO and Keck Science archives
271
+ (Baumgardt & Hilker 2018), and the dispersion from the integral-
272
+ field-unit data from the WAGGS project (Dalgleish et al. 2020). The
273
+ above data are expressed by open circles in Fig. 3. The data from the
274
+ MUSE survey (Kamann et al. 2018) were also used and denoted by
275
+ solid triangles. Some additional data were supplemented and marked
276
+ as crosses, such as those from McLaughlin et al. (2006) for NGC 104
277
+ and Larson & Seth (2015, private communication) for NGC 1851
278
+ and NGC 2808. (The data of McLaughlin et al. (2006) and Larson &
279
+ Seth (2015, private communication) were collected from the compi-
280
+ lation in Watkins et al. (2015b) and others were collected from the
281
+ compilation in the updated web catalog (third version) of Baumgardt
282
+ & Hilker (2018).)
283
+ For proper-motion velocity dispersion, we mainly took the data
284
+ from the Hubble Space Telescope from Watkins et al. (2015a) and the
285
+ Gaia data from Vasiliev & Baumgardt (2021). Open circles expressed
286
+ the former, and solid triangles expressed the latter in Fig. 4. Some
287
+ additional data were supplemented and denoted by crosses, which
288
+ include Häberle et al. (2021) for NGC 6441, McLaughlin et al. (2006)
289
+ for NGC 104, McNamara et al. (2003) for NGC 7078, McNamara
290
+ et al. (2012) for NGC 6266, and Zloczewski et al. (2012) for NGC
291
+ 6656 and NGC 6752. (The data of Vasiliev & Baumgardt (2021) and
292
+ Häberle et al. (2021) were collected from the updated web catalog of
293
+ Baumgardt & Hilker (2018), and the data of McLaughlin et al. (2006),
294
+ McNamara et al. (2003), McNamara et al. (2012), and Zloczewski
295
+ et al. (2012) were collected from Watkins et al. (2015b).)
296
+ Some proper motion data were downloaded in units of km/s,
297
+ which depends on the cluster distance written in the literature.
298
+ These data were transformed into mas/yr as the observational val-
299
+ ues for our work here. The transformation is 𝑣 = 𝑣0/𝐷𝐶, where 𝑣
300
+ and 𝑣0 are the velocity in mas/yr and km/s, 𝐷 is the distance and
301
+ 𝐶 = 4.74047 km yr kpc−1 mas−1 s−1 which is a factor for the unit
302
+ conversion (van Leeuwen 2009; Watkins et al. 2015b). The values
303
+ of cluster distances were taken from the corresponding literature. By
304
+ taking the root mean square of the upper and lower error bars from
305
+ the literature, we obtained a symmetric uncertainty for each data for
306
+ our work. Finally, to focus on the systems with enough observational
307
+ information, we studied 18 clusters with more than five data points
308
+ in each type of the above observational profiles.
309
+ 4 THE DETERMINATION OF PHYSICAL PARAMETERS
310
+ It was shown in Zocchi et al. (2017) that models with different
311
+ amounts of anisotropy could give the same surface brightness but
312
+ different kinematic profiles. Thus, using the surface brightness data
313
+ alone can lead to some degeneracy. Therefore, here we included
314
+ the surface brightness, the light-of-sight velocity dispersion, and the
315
+ proper-motion velocity dispersion data to obtain complete pictures of
316
+ the physical structures and kinematic properties of globular clusters
317
+ by determining related physical parameters through the data-model
318
+ fitting.
319
+ Following the method in Zocchi et al. (2017), we employed the
320
+ one-step fitting procedure with the single-mass LIMEPY models in this
321
+ paper. With all three considered types of observational data, a single
322
+ step of the fitting was performed to determine all cluster parameters.
323
+ The fitting was done through the minimization of the 𝜒2 function:
324
+ 𝜒2 = 𝜒2
325
+ sb + 𝜒2
326
+ los + 𝜒2
327
+ pm,
328
+ (5)
329
+ where 𝜒2
330
+ sb, 𝜒2
331
+ los, 𝜒2pm are the contributions from surface brightness,
332
+ line-of-sight velocity dispersion, and proper-motion velocity disper-
333
+ sion, respectively. They are defined by
334
+ 𝜒2
335
+ sb =
336
+ 𝑛sb
337
+ ∑︁
338
+ 𝑖=1
339
+ [𝜇𝑖 − ¯𝜇(𝑟𝑖)]2
340
+ 𝜖2
341
+ 𝜇,𝑖
342
+ ,
343
+ (6)
344
+ 𝜒2
345
+ los =
346
+ 𝑛los
347
+ ∑︁
348
+ 𝑖=1
349
+ [𝜎los,𝑖 − ¯𝜎los(𝑟����)]2
350
+ 𝜖2
351
+ los,𝑖
352
+ ,
353
+ (7)
354
+ and
355
+ 𝜒2
356
+ pm =
357
+ 𝑛pm
358
+ ∑︁
359
+ 𝑖=1
360
+ [𝜎pm,𝑖 − ¯𝜎pm(𝑟𝑖)]2
361
+ 𝜖2
362
+ pm,𝑖
363
+ ,
364
+ (8)
365
+ MNRAS 000, 1–15 (20XX)
366
+
367
+ 4
368
+ Cheng and Jiang
369
+ where 𝜇𝑖 is the 𝑖-th observational data of a surface brightness profile,
370
+ ¯𝜇(𝑟𝑖) is the theoretical surface brightness at that radial coordinate
371
+ 𝑟𝑖, and 𝜖𝜇,𝑖 is the error bar of the data 𝜇𝑖. Similarly, 𝜎los,𝑖, ¯𝜎los(𝑟𝑖),
372
+ 𝜖los,𝑖 are the corresponding quantities for line-of-sight velocity dis-
373
+ persion, and 𝜎pm,𝑖, ¯𝜎pm(𝑟𝑖), 𝜖pm,𝑖 are the observational data, the-
374
+ oretical value, and error bar for proper-motion velocity dispersion,
375
+ respectively. The numbers of observational data are 𝑛sb, 𝑛los, 𝑛pm,
376
+ respectively, for the surface brightness, line-of-sight velocity disper-
377
+ sion, and proper-motion velocity dispersion, individually.
378
+ The LIMEPY code was employed to obtain the above theoretical pro-
379
+ files. This code needed five input parameters, including the concen-
380
+ tration parameter 𝑊0, the truncation parameter 𝑔, the dimensionless
381
+ anisotropy radius ˆ𝑟a, the cluster mass 𝑀, and the half-mass radius
382
+ 𝑟h. The LIMEPY code generated several profiles, such as the surface
383
+ mass density Σ(𝑟𝑖), line-of-sight mean-square velocity 𝑢2
384
+ L(𝑟𝑖), radial
385
+ and tangential mean-square velocity on the projected plane 𝑢2
386
+ R(𝑟𝑖)
387
+ and 𝑢2
388
+ T(𝑟𝑖). Thus, the value of ¯𝜎los(𝑟𝑖) is simply the square root of
389
+ 𝑢2
390
+ L(𝑟𝑖), and ¯𝜎pm(𝑟𝑖) is the square root of [𝑢2
391
+ R(𝑟𝑖) + 𝑢2
392
+ T(𝑟𝑖)]/2.
393
+ To complete the data-model fitting, two more parameters were
394
+ needed. The cluster distance 𝐷 is a parameter that converts the radial
395
+ coordinate of the theoretical profile from pc to arcsec and the ob-
396
+ servational proper-motion velocity dispersion from mas/yr to km/s.
397
+ The V-band mass-to-light ratio Υ is a parameter for producing the
398
+ luminosity density Σ(𝑟𝑖)/Υ, and the surface brightness ¯𝜇(𝑟𝑖) can be
399
+ obtained by
400
+ ¯𝜇(𝑟𝑖) = 𝑀V,⊙ − 5(1 + log 𝑐) − 2.5 log(Σ(𝑟𝑖)/Υ),
401
+ (9)
402
+ where 𝑀V,⊙ = 4.83 mag is the V-band absolute magnitude of the
403
+ Sun and 𝑐 = 𝜋/648000 rad/arcsec is a factor for the unit conversion
404
+ (Watkins et al. 2015b).
405
+ Through the minimization of the 𝜒2 function, the best-fit values
406
+ of seven parameters 𝑊0, 𝑔, ˆ𝑟a, 𝑀, 𝑟h, 𝐷, Υ can be obtained. We
407
+ used the code EMCEE (Foreman-Mackey et al. 2013) to perform the 𝜒2
408
+ minimization. It is an affine-invariant ensemble sampler that employs
409
+ the Markov chain Monte Carlo (MCMC) process (Goodman & Weare
410
+ 2010). One has to decide the initial distribution and the parameters
411
+ range for the EMCEE samples. For the concentration parameter 𝑊0, the
412
+ range was set to 1 < 𝑊0 < 15. It covers a similar range in Table II
413
+ of King (1966) and represents various degrees of concentration of
414
+ star clusters. Figure 4 in Gieles & Zocchi (2015) showed the relevant
415
+ models for star clusters and the corresponding parameters; hence
416
+ we set 0 < 𝑔 < 3 for the truncation parameter accordingly. The
417
+ dimensionless anisotropy radius ˆ𝑟a needs a wide range to include the
418
+ isotropic models. Therefore, we set a large range for log ˆ𝑟a as −1 <
419
+ log ˆ𝑟a < 20. For the remained parameters, we checked the literature
420
+ values and considered wider ranges to include more possibilities. The
421
+ ranges of these parameters were set to be 0.1 < 𝑀 < 50 (105 M⊙),
422
+ 0.1 < 𝑟h < 15 (pc), 0.1 < 𝐷 < 35 (kpc), and 0.1 < Υ < 5 (Υ⊙).
423
+ Finally, the initial distributions of all parameters are set to be uniform.
424
+ 5 RESULTS AND DISCUSSION
425
+ The best-fit results are displayed in Table 1. The first column shows
426
+ the names of the clusters. Seven fitting parameters are listed from the
427
+ second to eighth columns. The second column presents the concen-
428
+ tration parameter 𝑊0 and the values range roughly from 3 to 9 for
429
+ these clusters. The third and the fourth columns show the truncation
430
+ parameter 𝑔 and the logarithm of the dimensionless anisotropy radius
431
+ log ˆ𝑟a. The fifth and sixth columns list the cluster mass 𝑀 and the
432
+ half-mass radius 𝑟h. These clusters have 𝑟h ≲ 10 pc. Among them,
433
+ NGC 5139 has the largest mass and radius. The heliocentric distance
434
+ 𝐷 is shown in the seventh column. Most clusters have 𝐷 ≲ 12 kpc
435
+ except for NGC 6715, which is roughly two times distant. The eighth
436
+ column reveals the V-band mass-to-light ratio Υ. To understand the
437
+ anisotropy conveniently, the quantity 𝜅 is shown in the ninth column.
438
+ NGC 5139 and NGC 7078 have 𝜅 > 1, indicating the anisotropic
439
+ behavior. The quantity in the last column is the reduced chi-square
440
+ 𝜒2r defined by
441
+ 𝜒2
442
+ r =
443
+ 𝜒2
444
+ 𝑛 − 𝑛p
445
+ ,
446
+ (10)
447
+ where 𝑛 is the total number of data and 𝑛p is the number of parame-
448
+ ters.
449
+ 5.1 Comparison with Previous Work
450
+ To compare our results with the previous work, we used the mea-
451
+ surable physical properties estimated in the published literature, as
452
+ listed in Table 2. We first considered the comparison of the cluster’s
453
+ total mass. In general, the masses estimated by Baumgardt & Hilker
454
+ (2018) are larger than those estimated by Watkins et al. (2015b), and
455
+ our results are usually between their values. Almost all of our results
456
+ are very close to the masses estimated in Watkins et al. (2015b).
457
+ We also compared our half-mass radius with the one in the catalog
458
+ of Baumgardt & Hilker (2018). Generally, our results are smaller,
459
+ consistent with the results of total mass, since our masses are lower
460
+ than those in Baumgardt & Hilker (2018). Therefore, the radii of the
461
+ clusters tend to be smaller to fit the line-of-sight velocity dispersion.
462
+ Some differences between the radius might come from the mass
463
+ spectrum. The radial distributions of different species may introduce
464
+ additional variation between the half-mass radii. Nevertheless, the
465
+ mass-to-light ratios obtained in our work are consistent with the
466
+ values in Baumgardt et al. (2020) and Watkins et al. (2015b).
467
+ For distance comparison, we compared with the values in Watkins
468
+ et al. (2015b), Baumgardt & Vasiliev (2021), and Harris (1996).
469
+ Watkins et al. (2015b) derived the distance by comparing their proper
470
+ motion velocity dispersion with the line-of-sight velocity dispersion
471
+ from the literature. Baumgardt & Vasiliev (2021) calculated the mean
472
+ distance from several methods, such as the Gaia EDR3 parallaxes,
473
+ the method by fitting nearby subdwarfs to globular cluster main
474
+ sequences, the color-magnitude diagram fitting, and the distances
475
+ from the period-luminosity relation of RR Lyrae stars. The distances
476
+ in Harris (1996) are a compilation of the distance measurements
477
+ from the literature.
478
+ Fig. 1 shows the ratio of our distance 𝐷 and the one published
479
+ in literature 𝐷lit, i.e., 𝐷/𝐷lit, for each considered cluster. For each
480
+ panel, the compared literature is labeled at the top-right corner. Each
481
+ point represents a particular cluster studied in the compared literature
482
+ and this work. The dashed line represents the unity, and the solid line
483
+ is the average value of the ratio. Two numbers are shown in the
484
+ bottom-right of the panels, the left number is the averaged 𝐷/𝐷lit,
485
+ and the right one is the averaged |𝐷/𝐷lit−1|. These numbers indicate
486
+ that our results are closer to Harris (1996) and Watkins et al. (2015b),
487
+ and slightly lower than Baumgardt & Vasiliev (2021). In general, our
488
+ results agree with the values from these studies.
489
+ 5.2 The Profiles
490
+ Fig. 2 to 4 show the profiles of surface brightness, line-of-sight veloc-
491
+ ity dispersion, and proper-motion velocity dispersion. The horizontal
492
+ axis is the distance from the cluster’s center in arcsec. The vertical
493
+ axis gives the surface brightness in mag/arcsec2 in Fig. 2, and ve-
494
+ locity dispersion in km/s from Fig. 3 to 4. It can be seen that LIMEPY
495
+ MNRAS 000, 1–15 (20XX)
496
+
497
+ Dynamical Properties of Globular Clusters
498
+ 5
499
+ Table 1. The properties of the clusters. The first column lists the names of the clusters. Columns two to eight show the fitting parameters, which are concentration
500
+ parameter 𝑊0, truncation parameter 𝑔, the logarithm of the dimensionless anisotropy radius log ˆ𝑟a, cluster mass 𝑀, half-mass radius 𝑟h, distance 𝐷, and V-band
501
+ mass-to-light ratio Υ. Column nine presents the quantity 𝜅 which measures the amount of anisotropy, and the final column gives 𝜒2r .
502
+ cluster
503
+ 𝑊0
504
+ 𝑔
505
+ log ˆ𝑟a
506
+ 𝑀
507
+ 𝑟h
508
+ 𝐷
509
+ Υ
510
+ 𝜅
511
+ 𝜒2r
512
+ (105 M⊙)
513
+ (pc)
514
+ (kpc)
515
+ (Υ⊙)
516
+ NGC 104
517
+ 8.36 ± 0.06
518
+ 1.31 ± 0.03
519
+ 11.13+6.02
520
+ −6.10
521
+ 6.87 ± 0.15
522
+ 5.21 ± 0.12
523
+ 4.33 ± 0.03
524
+ 1.53 ± 0.03
525
+ 1.00
526
+ 2.10
527
+ NGC 288
528
+ 4.46+0.47
529
+ −0.82
530
+ 1.55+0.52
531
+ −0.38
532
+ 10.40+6.44
533
+ −6.41
534
+ 1.02+0.11
535
+ −0.10
536
+ 8.26+0.33
537
+ −0.32
538
+ 9.80+0.37
539
+ −0.36
540
+ 2.32 ± 0.12
541
+ 1.00
542
+ 1.18
543
+ NGC 362
544
+ 7.20 ± 0.10
545
+ 1.67 ± 0.06
546
+ 11.24+5.87
547
+ −6.48
548
+ 2.09+0.11
549
+ −0.10
550
+ 2.36+0.08
551
+ −0.07
552
+ 8.71+0.16
553
+ −0.15
554
+ 1.22 ± 0.03
555
+ 1.00
556
+ 4.74
557
+ NGC 1851
558
+ 7.33+0.19
559
+ −0.20
560
+ 2.04 ± 0.09
561
+ 10.77+6.26
562
+ −6.12
563
+ 2.28+0.10
564
+ −0.09
565
+ 2.15+0.15
566
+ −0.13
567
+ 10.82+0.15
568
+ −0.14
569
+ 1.73+0.09
570
+ −0.08
571
+ 1.00
572
+ 1.61
573
+ NGC 2808
574
+ 6.27+0.17
575
+ −0.16
576
+ 2.02+0.10
577
+ −0.07
578
+ 11.81+5.01
579
+ −6.78
580
+ 6.57+0.25
581
+ −0.19
582
+ 2.69+0.08
583
+ −0.06
584
+ 9.63+0.12
585
+ −0.10
586
+ 1.56+0.05
587
+ −0.04
588
+ 1.00
589
+ 1.63
590
+ NGC 3201
591
+ 5.89+0.31
592
+ −0.34
593
+ 2.45 ± 0.09
594
+ 11.00+6.19
595
+ −6.29
596
+ 1.21+0.08
597
+ −0.07
598
+ 5.21+0.42
599
+ −0.33
600
+ 4.38+0.10
601
+ −0.09
602
+ 2.33+0.12
603
+ −0.11
604
+ 1.00
605
+ 2.74
606
+ NGC 5139
607
+ 4.02+0.48
608
+ −0.65
609
+ 1.94+0.27
610
+ −0.26
611
+ 0.41+0.08
612
+ −0.10
613
+ 32.82+0.65
614
+ −0.67
615
+ 8.82+0.19
616
+ −0.17
617
+ 5.32 ± 0.03
618
+ 2.38 ± 0.09
619
+ 1.15
620
+ 3.86
621
+ NGC 5904
622
+ 7.03+0.09
623
+ −0.10
624
+ 1.56+0.05
625
+ −0.04
626
+ 10.39+6.55
627
+ −6.07
628
+ 3.03+0.16
629
+ −0.15
630
+ 4.54+0.12
631
+ −0.11
632
+ 7.24 ± 0.13
633
+ 1.39+0.04
634
+ −0.03
635
+ 1.00
636
+ 1.85
637
+ NGC 6121
638
+ 7.52+0.16
639
+ −0.13
640
+ 0.46+0.31
641
+ −0.21
642
+ 9.80+6.94
643
+ −5.59
644
+ 0.81+0.05
645
+ −0.04
646
+ 3.20+0.17
647
+ −0.13
648
+ 1.85+0.04
649
+ −0.03
650
+ 2.11+0.10
651
+ −0.08
652
+ 1.00
653
+ 1.12
654
+ NGC 6218
655
+ 5.77+0.29
656
+ −0.35
657
+ 1.51+0.25
658
+ −0.22
659
+ 10.69+6.36
660
+ −6.53
661
+ 0.75 ± 0.06
662
+ 3.02+0.14
663
+ −0.13
664
+ 4.59+0.15
665
+ −0.14
666
+ 1.78+0.09
667
+ −0.08
668
+ 1.00
669
+ 1.19
670
+ NGC 6266
671
+ 7.84+0.08
672
+ −0.09
673
+ 0.62+0.11
674
+ −0.10
675
+ 10.90 ± 6.23
676
+ 5.98+0.25
677
+ −0.24
678
+ 2.55 ± 0.08
679
+ 6.33+0.09
680
+ −0.08
681
+ 1.85 ± 0.05
682
+ 1.00
683
+ 1.57
684
+ NGC 6388
685
+ 7.09+0.10
686
+ −0.11
687
+ 1.68+0.09
688
+ −0.08
689
+ 10.86+6.16
690
+ −6.13
691
+ 7.79 ± 0.20
692
+ 2.07 ± 0.05
693
+ 10.35 ± 0.10
694
+ 1.68 ± 0.03
695
+ 1.00
696
+ 2.94
697
+ NGC 6397
698
+ 9.17 ± 0.17
699
+ 0.87 ± 0.08
700
+ 10.96+6.13
701
+ −6.11
702
+ 0.79+0.04
703
+ −0.03
704
+ 3.73 ± 0.19
705
+ 2.40 ± 0.04
706
+ 2.47 ± 0.12
707
+ 1.00
708
+ 1.97
709
+ NGC 6441
710
+ 7.75 ± 0.06
711
+ 1.24+0.10
712
+ −0.09
713
+ 10.74+6.31
714
+ −6.27
715
+ 10.54+0.28
716
+ −0.27
717
+ 2.90+0.07
718
+ −0.06
719
+ 11.91 ± 0.11
720
+ 1.82 ± 0.03
721
+ 1.00
722
+ 3.35
723
+ NGC 6656
724
+ 6.48+0.23
725
+ −0.26
726
+ 1.87+0.34
727
+ −0.39
728
+ 10.73+6.33
729
+ −6.50
730
+ 3.57+0.22
731
+ −0.20
732
+ 4.40+0.29
733
+ −0.20
734
+ 3.10 ± 0.05
735
+ 1.85 ± 0.07
736
+ 1.00
737
+ 1.04
738
+ NGC 6715
739
+ 6.99 ± 0.07
740
+ 2.21+0.02
741
+ −0.03
742
+ 11.26+6.00
743
+ −6.28
744
+ 17.79+1.12
745
+ −1.06
746
+ 5.28+0.29
747
+ −0.25
748
+ 25.08 ± 0.53
749
+ 2.07 ± 0.06
750
+ 1.00
751
+ 2.83
752
+ NGC 6752
753
+ 8.35+0.12
754
+ −0.13
755
+ 1.38 ± 0.06
756
+ 10.95+6.16
757
+ −6.21
758
+ 1.92 ± 0.09
759
+ 3.45 ± 0.16
760
+ 4.13 ± 0.06
761
+ 2.24 ± 0.08
762
+ 1.00
763
+ 1.20
764
+ NGC 7078
765
+ 8.30+0.12
766
+ −0.13
767
+ 0.86+0.15
768
+ −0.13
769
+ 1.16+0.07
770
+ −0.06
771
+ 5.08 ± 0.17
772
+ 4.05+0.18
773
+ −0.17
774
+ 10.40 ± 0.12
775
+ 1.53 ± 0.05
776
+ 1.16
777
+ 1.46
778
+ Table 2. The literature parameters of the clusters. The number in the parentheses represents the literature, (1) stands for Baumgardt & Hilker (2018), (2) refers
779
+ to Watkins et al. (2015b), (3) corresponds to Baumgardt & Vasiliev (2021), (4) represents Harris (1996), and (5) is Baumgardt et al. (2020). The updated values
780
+ for (1) and (5) are picked from the web catalog of Baumgardt & Hilker (2018).
781
+ cluster
782
+ 𝑀
783
+ 𝑀
784
+ 𝑟h
785
+ 𝐷
786
+ 𝐷
787
+ 𝐷
788
+ Υ
789
+ Υ
790
+ (105 M⊙)
791
+ (105 M⊙)
792
+ (pc)
793
+ (kpc)
794
+ (kpc)
795
+ (kpc)
796
+ (Υ⊙)
797
+ (Υ⊙)
798
+ (1)
799
+ (2)
800
+ (1)
801
+ (2)
802
+ (3)
803
+ (4)
804
+ (5)
805
+ (2)
806
+ NGC 104
807
+ 8.95 ± 0.06
808
+ 5.57+0.33
809
+ −0.28
810
+ 6.30
811
+ 4.15 ± 0.08
812
+ 4.521 ± 0.031
813
+ 4.5
814
+ 1.96 ± 0.09
815
+ 1.40 ± 0.03
816
+ NGC 288
817
+ 0.934 ± 0.026
818
+ 0.79+0.13
819
+ −0.11
820
+ 8.37
821
+ 9.03+0.48
822
+ −0.56
823
+ 8.988+0.089
824
+ −0.088
825
+ 8.9
826
+ 2.16 ± 0.10
827
+ 2.20+0.13
828
+ −0.10
829
+ NGC 362
830
+ 2.84 ± 0.04
831
+ ...
832
+ 3.79
833
+ ...
834
+ 8.829 ± 0.096
835
+ 8.6
836
+ 1.44 ± 0.05
837
+ ...
838
+ NGC 1851
839
+ 3.18 ± 0.04
840
+ 1.78+0.10
841
+ −0.11
842
+ 2.90
843
+ 10.32+0.20
844
+ −0.24
845
+ 11.951+0.134
846
+ −0.133
847
+ 12.1
848
+ 1.66 ± 0.06
849
+ 1.51 ± 0.03
850
+ NGC 2808
851
+ 8.64 ± 0.06
852
+ 5.91+0.22
853
+ −0.25
854
+ 3.89
855
+ 9.45+0.13
856
+ −0.15
857
+ 10.060+0.112
858
+ −0.111
859
+ 9.6
860
+ 1.51 ± 0.06
861
+ 1.56 ± 0.02
862
+ NGC 3201
863
+ 1.60 ± 0.03
864
+ ...
865
+ 6.78
866
+ ...
867
+ 4.737+0.043
868
+ −0.042
869
+ 4.9
870
+ 2.16 ± 0.09
871
+ ...
872
+ NGC 5139
873
+ 36.4 ± 0.4
874
+ 34.52+1.45
875
+ −1.43
876
+ 10.36
877
+ 5.19+0.07
878
+ −0.08
879
+ 5.426 ± 0.047
880
+ 5.2
881
+ 2.58 ± 0.10
882
+ 2.66 ± 0.04
883
+ NGC 5904
884
+ 3.94 ± 0.06
885
+ 3.65 ± 0.75
886
+ 5.68
887
+ 7.79+0.47
888
+ −0.61
889
+ 7.479 ± 0.060
890
+ 7.5
891
+ 1.81 ± 0.06
892
+ 1.43+0.09
893
+ −0.10
894
+ NGC 6121
895
+ 0.871 ± 0.011
896
+ ...
897
+ 3.69
898
+ ...
899
+ 1.851+0.015
900
+ −0.016
901
+ 2.2
902
+ 1.59 ± 0.06
903
+ ...
904
+ NGC 6218
905
+ 1.07 ± 0.03
906
+ ...
907
+ 4.05
908
+ ...
909
+ 5.109+0.049
910
+ −0.048
911
+ 4.8
912
+ 1.92 ± 0.09
913
+ ...
914
+ NGC 6266
915
+ 6.10 ± 0.04
916
+ 6.09+0.39
917
+ −0.33
918
+ 2.43
919
+ 6.42 ± 0.14
920
+ 6.412+0.105
921
+ −0.104
922
+ 6.8
923
+ 1.99 ± 0.11
924
+ 2.22 ± 0.04
925
+ NGC 6388
926
+ 12.5 ± 0.1
927
+ 8.27+0.89
928
+ −0.95
929
+ 4.34
930
+ 10.90+0.40
931
+ −0.45
932
+ 11.171+0.162
933
+ −0.161
934
+ 9.9
935
+ 2.19 ± 0.06
936
+ 1.68+0.06
937
+ −0.07
938
+ NGC 6397
939
+ 0.966 ± 0.013
940
+ 0.70+0.09
941
+ −0.08
942
+ 3.90
943
+ 2.39+0.13
944
+ −0.11
945
+ 2.482 ± 0.019
946
+ 2.3
947
+ 1.66 ± 0.07
948
+ 2.23+0.10
949
+ −0.09
950
+ NGC 6441
951
+ 13.2 ± 0.1
952
+ ...
953
+ 3.47
954
+ ...
955
+ 12.728+0.163
956
+ −0.162
957
+ 11.6
958
+ 1.77 ± 0.13
959
+ ...
960
+ NGC 6656
961
+ 4.76 ± 0.05
962
+ 2.49+0.44
963
+ −0.37
964
+ 5.29
965
+ 2.84 ± 0.16
966
+ 3.303 ± 0.037
967
+ 3.2
968
+ 2.05 ± 0.08
969
+ 1.88+0.12
970
+ −0.10
971
+ NGC 6715
972
+ 17.8 ± 0.3
973
+ 11.83+0.62
974
+ −0.53
975
+ 5.20
976
+ 22.57+0.44
977
+ −0.39
978
+ 26.283+0.328
979
+ −0.325
980
+ 26.5
981
+ 2.10 ± 0.12
982
+ 1.94 ± 0.03
983
+ NGC 6752
984
+ 2.76 ± 0.04
985
+ 1.82 ± 0.12
986
+ 5.27
987
+ 4.02+0.10
988
+ −0.08
989
+ 4.125 ± 0.041
990
+ 4.0
991
+ 2.34 ± 0.11
992
+ 2.14+0.05
993
+ −0.06
994
+ NGC 7078
995
+ 6.33 ± 0.07
996
+ 4.95 ± 0.19
997
+ 4.30
998
+ 10.36+0.15
999
+ −0.16
1000
+ 10.709+0.096
1001
+ −0.095
1002
+ 10.4
1003
+ 1.58 ± 0.10
1004
+ 1.49 ± 0.02
1005
+ MNRAS 000, 1–15 (20XX)
1006
+
1007
+ 6
1008
+ Cheng and Jiang
1009
+ 5
1010
+ 10
1011
+ 15
1012
+ 20
1013
+ 25
1014
+ D [kpc]
1015
+ 0.8
1016
+ 0.9
1017
+ 1.0
1018
+ 1.1
1019
+ 1.2
1020
+ D / Dlit
1021
+ Watkins et al. (2015b)
1022
+ 1.02, 0.05
1023
+ 5
1024
+ 10
1025
+ 15
1026
+ 20
1027
+ 25
1028
+ D [kpc]
1029
+ 0.8
1030
+ 0.9
1031
+ 1.0
1032
+ 1.1
1033
+ 1.2
1034
+ Baumgardt & Vasiliev (2021)
1035
+ 0.96, 0.05
1036
+ 5
1037
+ 10
1038
+ 15
1039
+ 20
1040
+ 25
1041
+ D [kpc]
1042
+ 0.8
1043
+ 0.9
1044
+ 1.0
1045
+ 1.1
1046
+ 1.2
1047
+ Harris (1996)
1048
+ 0.98, 0.05
1049
+ Figure 1. The comparison of the cluster distance with the values mentioned in earlier studies. The horizontal axis is the cluster distance obtained in this work,
1050
+ and the vertical axis shows the ratio of our value to the distance given in the literature. The dashed line and the solid line represent the unity and the average. Each
1051
+ panel is for comparison with the particular publication, as labeled at the top-right corner. At the bottom-right corner, the left number is the averaged 𝐷/𝐷lit,
1052
+ and the right number is the averaged |𝐷/𝐷lit − 1|.
1053
+ models can produce similar profiles as observational ones. To ex-
1054
+ amine these clusters more quantitatively, we classified the results by
1055
+ 𝜒2r . Many clusters were found to have 𝜒2r < 2. These clusters have
1056
+ suitable fittings for all three profiles, as shown in the figures.
1057
+ NGC 362 has the largest 𝜒2r , and the model profiles agree with the
1058
+ observations in surface brightness and line-of-sight velocity disper-
1059
+ sion. However, the central part of the modeled proper-motion velocity
1060
+ dispersion is slightly larger than the observations. Data with a small
1061
+ error bar in the outer part located much higher than the profile, mak-
1062
+ ing the fitting worse. NGC 6441 also has a larger 𝜒2r . The model
1063
+ agrees well with the surface brightness and the outer part of the
1064
+ proper motion velocity dispersion but predicts larger values for the
1065
+ inner part. The model can also fit the rough trend of the line-of-sight
1066
+ velocity dispersion, but some points lie below the model.
1067
+ For NGC 3201, the model has smaller line-of-sight velocity dis-
1068
+ persion for radius above 100 arcsec. There are also some under
1069
+ estimations for the proper motions in the outermost region, where
1070
+ the observational profile tends to level off rather than continue to
1071
+ decrease. Some scenarios were proposed to explain the higher ve-
1072
+ locity dispersion in the outer part, such as the orbital history with
1073
+ accretion and the embedding by a dark matter halo (Bianchini et al.
1074
+ 2019). It was also found that binary stars could contribute to part of
1075
+ the effect (Wan et al. 2021). For NGC 6715, the model agrees with
1076
+ the observations, except for the outermost region of the line-of-sight
1077
+ velocity dispersion, where the observational profile grows. This rise
1078
+ is probably caused by the stars in the nucleus of the Sagittarius dwarf
1079
+ galaxy, where NGC 6715 inhabits (Bellazzini et al. 2008).
1080
+ NGC 5139 has large central velocity dispersions, which the model
1081
+ cannot explain well. For NGC 6388, the model has a steeper proper-
1082
+ motion velocity dispersion profile than the observational one. Further
1083
+ discussions of these two clusters will be made in the following sub-
1084
+ section.
1085
+ 5.3 Possible Intermediate-Mass Black Hole ?
1086
+ Stellar black holes exist in astrophysical systems such as X-ray bina-
1087
+ ries (Mikolajewska et al. 2022). In addition, supermassive black holes
1088
+ are also confirmed to exist at the centers of our Milky Way (GRAV-
1089
+ ITY Collaboration et al. 2019) and other galaxies (Blandford et al.
1090
+ 2019). Whether there are any intermediate-mass black holes in the
1091
+ universe is one of the most important questions in astronomy. Globu-
1092
+ lar clusters are considered good candidates to host intermediate-mass
1093
+ black holes and thus attract much attention. Among 18 globular clus-
1094
+ ters in the present work, NGC 5139 was discussed previously as a
1095
+ likely candidate.
1096
+ For our work here, the data-model fitting of NGC 5139 led to
1097
+ two groups of model parameters, as shown in Fig. 5. These groups
1098
+ have very different concentration parameters 𝑊0 and logarithm of
1099
+ the dimensionless anisotropy radius log ˆ𝑟a. One has smaller 𝑊0 and
1100
+ log ˆ𝑟a, and the other has larger values. Hence, we do further fittings
1101
+ with narrower ranges as 1 < 𝑊0 < 8, −1 < log ˆ𝑟a < 2, and 8 < 𝑊0 <
1102
+ 15, 2 < log ˆ𝑟a < 20, separately. The results are shown in Table 3.
1103
+ We denote the one with lower 𝜒2r as Model A, the result previously
1104
+ listed in Table 1 and presented in Fig. 2 to 4. Model A has a low
1105
+ concentration. It also has a small dimensionless anisotropy radius
1106
+ with 𝜅 = 1.15, making it more anisotropic. In contrast, Model B is
1107
+ isotropic with a high concentration.
1108
+ The best-fit profiles are shown in Fig. 6. Model A fits the surface
1109
+ brightness well but predicts lower central velocity dispersion, espe-
1110
+ cially for the proper motion. On the other hand, Model B has good
1111
+ fittings on both velocity dispersion but a poor fitting on the surface
1112
+ brightness. The deviation in surface brightness leads to a larger 𝜒2r .
1113
+ Although the data and radial range of the observational kinematic
1114
+ profiles differs, the parameters from Model A agree with those in the
1115
+ best-fit model in Zocchi et al. (2017).
1116
+ These results obtained with two models show that it is difficult
1117
+ to perfectly and simultaneously fit all profiles of NGC 5139 with
1118
+ the current considered model. This could indicate the existence of
1119
+ central dark objects which can cause an increase in central velocities.
1120
+ These objects could be an intermediate-mass black hole (Noyola et al.
1121
+ 2010; Baumgardt 2017) or a group of stellar-mass black holes at the
1122
+ cluster center (Baumgardt et al. 2019b). Both can also suppress the
1123
+ mass segregation of the stars (Gill et al. 2008; Peuten et al. 2016)
1124
+ and render the cluster to have a larger core (Baumgardt et al. 2005;
1125
+ Peuten et al. 2017). The main difference is that the intermediate-
1126
+ mass black hole could produce some stars faster than 60 km/s in the
1127
+ central 20 arcsec of NGC 5139, which was not confirmed in current
1128
+ observations (Baumgardt et al. 2019b).
1129
+ NGC 6388 is another candidate cluster that may host a central
1130
+ intermediate-mass black hole. The study of the integrated light spec-
1131
+ tra revealed a high central LOS velocity dispersion ∼25 km/s within 2
1132
+ arcsec (Lützgendorf et al. 2011). However, there was also a result that
1133
+ suggests a dispersion ∼15 km/s in the same region derived from stars’
1134
+ radial velocities (Lanzoni et al. 2013). Hence, the actual kinematic
1135
+ behavior of the cluster center is not clear. The data we used have
1136
+ the extension to nearly 5 arcsec with a velocity dispersion ∼20 km/s.
1137
+ Our results show that the surface brightness and line-of-sight velocity
1138
+ MNRAS 000, 1–15 (20XX)
1139
+
1140
+ Dynamical Properties of Globular Clusters
1141
+ 7
1142
+ 100
1143
+ 101
1144
+ 102
1145
+ 103
1146
+ 15
1147
+ 20
1148
+ 25
1149
+ μ [mag/arcsec2]
1150
+ NGC 104
1151
+ 100
1152
+ 101
1153
+ 102
1154
+ 103
1155
+ 20
1156
+ 25
1157
+ 30
1158
+ NGC 288
1159
+ 100
1160
+ 101
1161
+ 102
1162
+ 103
1163
+ 15
1164
+ 20
1165
+ 25
1166
+ NGC 362
1167
+ 100
1168
+ 101
1169
+ 102
1170
+ 103
1171
+ 15
1172
+ 20
1173
+ 25
1174
+ μ [mag/arcsec2]
1175
+ NGC 1851
1176
+ 100
1177
+ 101
1178
+ 102
1179
+ 103
1180
+ 15
1181
+ 20
1182
+ 25
1183
+ NGC 2808
1184
+ 100
1185
+ 101
1186
+ 102
1187
+ 103
1188
+ 20
1189
+ 25
1190
+ NGC 3201
1191
+ 101
1192
+ 102
1193
+ 103
1194
+ 15
1195
+ 20
1196
+ 25
1197
+ μ [mag/arcsec2]
1198
+ NGC 5139
1199
+ 101
1200
+ 102
1201
+ 103
1202
+ 15
1203
+ 20
1204
+ 25
1205
+ 30
1206
+ NGC 5904
1207
+ 100
1208
+ 101
1209
+ 102
1210
+ 103
1211
+ 15
1212
+ 20
1213
+ 25
1214
+ NGC 6121
1215
+ 100
1216
+ 101
1217
+ 102
1218
+ 103
1219
+ 15
1220
+ 20
1221
+ 25
1222
+ μ [mag/arcsec2]
1223
+ NGC 6218
1224
+ 100
1225
+ 101
1226
+ 102
1227
+ 103
1228
+ 15
1229
+ 20
1230
+ NGC 6266
1231
+ 100
1232
+ 101
1233
+ 102
1234
+ 15
1235
+ 20
1236
+ 25
1237
+ NGC 6388
1238
+ 100
1239
+ 101
1240
+ 102
1241
+ 103
1242
+ 15
1243
+ 20
1244
+ 25
1245
+ μ [mag/arcsec2]
1246
+ NGC 6397
1247
+ 100
1248
+ 101
1249
+ 102
1250
+ 12.5
1251
+ 15.0
1252
+ 17.5
1253
+ 20.0
1254
+ 22.5
1255
+ NGC 6441
1256
+ 100
1257
+ 101
1258
+ 102
1259
+ 103
1260
+ 15.0
1261
+ 17.5
1262
+ 20.0
1263
+ 22.5
1264
+ 25.0
1265
+ NGC 6656
1266
+ 100
1267
+ 101
1268
+ 102
1269
+ 103
1270
+ r [arcsec]
1271
+ 15
1272
+ 20
1273
+ 25
1274
+ μ [mag/arcsec2]
1275
+ NGC 6715
1276
+ 100
1277
+ 101
1278
+ 102
1279
+ 103
1280
+ r [arcsec]
1281
+ 15
1282
+ 20
1283
+ 25
1284
+ NGC 6752
1285
+ 100
1286
+ 101
1287
+ 102
1288
+ 103
1289
+ r [arcsec]
1290
+ 15
1291
+ 20
1292
+ 25
1293
+ NGC 7078
1294
+ Figure 2. The surface brightness profiles of the clusters. The observations are shown as crosses, and the models are expressed by grey lines. For each panel, the
1295
+ name of the cluster is mentioned at the top-right corner.
1296
+ MNRAS 000, 1–15 (20XX)
1297
+
1298
+ 8
1299
+ Cheng and Jiang
1300
+ 101
1301
+ 102
1302
+ 103
1303
+ 0
1304
+ 5
1305
+ 10
1306
+ 15
1307
+ σlos [km/s]
1308
+ NGC 104
1309
+ 102
1310
+ 103
1311
+ 0
1312
+ 1
1313
+ 2
1314
+ 3
1315
+ 4
1316
+ NGC 288
1317
+ 101
1318
+ 102
1319
+ 103
1320
+ 0.0
1321
+ 2.5
1322
+ 5.0
1323
+ 7.5
1324
+ 10.0
1325
+ NGC 362
1326
+ 101
1327
+ 102
1328
+ 103
1329
+ 0
1330
+ 5
1331
+ 10
1332
+ σlos [km/s]
1333
+ NGC 1851
1334
+ 101
1335
+ 102
1336
+ 0
1337
+ 5
1338
+ 10
1339
+ 15
1340
+ NGC 2808
1341
+ 101
1342
+ 102
1343
+ 103
1344
+ 0
1345
+ 2
1346
+ 4
1347
+ NGC 3201
1348
+ 101
1349
+ 102
1350
+ 103
1351
+ 0
1352
+ 10
1353
+ 20
1354
+ 30
1355
+ σlos [km/s]
1356
+ NGC 5139
1357
+ 101
1358
+ 102
1359
+ 103
1360
+ 0.0
1361
+ 2.5
1362
+ 5.0
1363
+ 7.5
1364
+ 10.0
1365
+ NGC 5904
1366
+ 102
1367
+ 103
1368
+ 0
1369
+ 2
1370
+ 4
1371
+ 6
1372
+ NGC 6121
1373
+ 101
1374
+ 102
1375
+ 0
1376
+ 2
1377
+ 4
1378
+ σlos [km/s]
1379
+ NGC 6218
1380
+ 101
1381
+ 102
1382
+ 0
1383
+ 5
1384
+ 10
1385
+ 15
1386
+ 20
1387
+ NGC 6266
1388
+ 101
1389
+ 102
1390
+ 0
1391
+ 10
1392
+ 20
1393
+ NGC 6388
1394
+ 101
1395
+ 102
1396
+ 103
1397
+ 0
1398
+ 2
1399
+ 4
1400
+ 6
1401
+ σlos [km/s]
1402
+ NGC 6397
1403
+ 101
1404
+ 102
1405
+ 0
1406
+ 10
1407
+ 20
1408
+ NGC 6441
1409
+ 101
1410
+ 102
1411
+ 103
1412
+ 0
1413
+ 5
1414
+ 10
1415
+ NGC 6656
1416
+ 101
1417
+ 102
1418
+ 103
1419
+ r [arcsec]
1420
+ 0
1421
+ 5
1422
+ 10
1423
+ 15
1424
+ 20
1425
+ σlos [km/s]
1426
+ NGC 6715
1427
+ 101
1428
+ 102
1429
+ 103
1430
+ r [arcsec]
1431
+ 0.0
1432
+ 2.5
1433
+ 5.0
1434
+ 7.5
1435
+ 10.0
1436
+ NGC 6752
1437
+ 101
1438
+ 102
1439
+ 103
1440
+ r [arcsec]
1441
+ 0
1442
+ 5
1443
+ 10
1444
+ 15
1445
+ NGC 7078
1446
+ Figure 3. The line-of-sight velocity dispersion profiles of the clusters. The open circles represent the data of Baumgardt (2017), Baumgardt & Hilker (2018),
1447
+ and Dalgleish et al. (2020). The data of Kamann et al. (2018) are shown in solid triangles. The crosses are used for additional data of some clusters mentioned
1448
+ in Section 3. The models are expressed by grey lines. For each panel, the name of the cluster is mentioned at the top-right corner.
1449
+ MNRAS 000, 1–15 (20XX)
1450
+
1451
+ Dynamical Properties of Globular Clusters
1452
+ 9
1453
+ 101
1454
+ 102
1455
+ 103
1456
+ 0
1457
+ 5
1458
+ 10
1459
+ 15
1460
+ σpm [km/s]
1461
+ NGC 104
1462
+ 101
1463
+ 102
1464
+ 0
1465
+ 2
1466
+ 4
1467
+ NGC 288
1468
+ 101
1469
+ 102
1470
+ 0.0
1471
+ 2.5
1472
+ 5.0
1473
+ 7.5
1474
+ 10.0
1475
+ NGC 362
1476
+ 101
1477
+ 102
1478
+ 103
1479
+ 0.0
1480
+ 2.5
1481
+ 5.0
1482
+ 7.5
1483
+ 10.0
1484
+ σpm [km/s]
1485
+ NGC 1851
1486
+ 101
1487
+ 102
1488
+ 103
1489
+ 0
1490
+ 5
1491
+ 10
1492
+ 15
1493
+ NGC 2808
1494
+ 102
1495
+ 103
1496
+ 0
1497
+ 1
1498
+ 2
1499
+ 3
1500
+ 4
1501
+ NGC 3201
1502
+ 101
1503
+ 102
1504
+ 103
1505
+ 0
1506
+ 10
1507
+ 20
1508
+ 30
1509
+ σpm [km/s]
1510
+ NGC 5139
1511
+ 100
1512
+ 101
1513
+ 102
1514
+ 0.0
1515
+ 2.5
1516
+ 5.0
1517
+ 7.5
1518
+ 10.0
1519
+ NGC 5904
1520
+ 102
1521
+ 103
1522
+ 0
1523
+ 2
1524
+ 4
1525
+ NGC 6121
1526
+ 102
1527
+ 103
1528
+ 0
1529
+ 1
1530
+ 2
1531
+ 3
1532
+ 4
1533
+ σpm [km/s]
1534
+ NGC 6218
1535
+ 100
1536
+ 101
1537
+ 102
1538
+ 0
1539
+ 10
1540
+ 20
1541
+ NGC 6266
1542
+ 101
1543
+ 102
1544
+ 0
1545
+ 5
1546
+ 10
1547
+ 15
1548
+ 20
1549
+ NGC 6388
1550
+ 101
1551
+ 102
1552
+ 103
1553
+ 0
1554
+ 2
1555
+ 4
1556
+ 6
1557
+ σpm [km/s]
1558
+ NGC 6397
1559
+ 100
1560
+ 101
1561
+ 102
1562
+ 0
1563
+ 10
1564
+ 20
1565
+ NGC 6441
1566
+ 101
1567
+ 102
1568
+ 103
1569
+ 0.0
1570
+ 2.5
1571
+ 5.0
1572
+ 7.5
1573
+ 10.0
1574
+ NGC 6656
1575
+ 101
1576
+ 102
1577
+ r [arcsec]
1578
+ 0
1579
+ 5
1580
+ 10
1581
+ 15
1582
+ 20
1583
+ σpm [km/s]
1584
+ NGC 6715
1585
+ 101
1586
+ 102
1587
+ 103
1588
+ r [arcsec]
1589
+ 0.0
1590
+ 2.5
1591
+ 5.0
1592
+ 7.5
1593
+ 10.0
1594
+ NGC 6752
1595
+ 100
1596
+ 101
1597
+ 102
1598
+ r [arcsec]
1599
+ 0
1600
+ 5
1601
+ 10
1602
+ 15
1603
+ 20
1604
+ NGC 7078
1605
+ Figure 4. The proper-motion velocity dispersion profiles of the clusters. The open circles represent the data of Watkins et al. (2015a). The data of Vasiliev &
1606
+ Baumgardt (2021) is shown in solid triangles. The crosses are used for additional data of some clusters mentioned in Section 3. The models are expressed by
1607
+ grey lines. For each panel, the name of the cluster is mentioned at the top-right corner.
1608
+ MNRAS 000, 1–15 (20XX)
1609
+
1610
+ 10
1611
+ Cheng and Jiang
1612
+ dispersion can be fitted well without the central black hole. However,
1613
+ the model predicts a steeper proper-motion velocity-dispersion pro-
1614
+ file than the observations, being higher inside but lower outside. This
1615
+ behavior can also be seen in Figure 9 of Watkins et al. (2015b).
1616
+ NGC 7078 is also a candidate cluster that could host an
1617
+ intermediate-mass black hole. The increase in central velocity dis-
1618
+ persion found in Hubble Space Telescope was explained by an
1619
+ intermediate-mass black hole (Gerssen et al. 2002). However, the
1620
+ cluster can also be fitted with a group of dark stellar remnants (den
1621
+ Brok et al. 2014) or N-body simulations without intermediate-mass
1622
+ black holes (Baumgardt 2017). In our results, the cluster could be fit-
1623
+ ted well without central black holes, and some degree of anisotropy
1624
+ was observed, which can raise the central velocities. In addition,
1625
+ although there are raised velocity dispersions in observation, the un-
1626
+ certainties of the data are also large. Therefore, we obtain a better
1627
+ fitting than NGC 5139.
1628
+ 5.4 The Anisotropy
1629
+ Two clusters, NGC 5139 and NGC 7078, possess small dimensionless
1630
+ anisotropy radius and reveal some degree of anisotropy. The former
1631
+ has 𝜅 = 1.15 and the latter has 𝜅 = 1.16. Other clusters have isotropic
1632
+ behavior with 𝜅 = 1.00 and a large anisotropy radius. One effect of
1633
+ radial anisotropy is that it can increase the central velocity dispersion.
1634
+ The rise in central velocity dispersions can be seen in Fig. 3 and 4. On
1635
+ the other hand, the amount of anisotropy estimated from our fittings
1636
+ could be underestimated, since the difference between tangential and
1637
+ radial proper motions will be averaged out in the combined proper
1638
+ motion velocity dispersion.
1639
+ The results are reasonable compared with some previous studies.
1640
+ For example, the parameters of NGC 5139 are similar to those es-
1641
+ timated in Zocchi et al. (2017) which the fittings were carried out
1642
+ with both radial and tangential proper motion velocity dispersions.
1643
+ The weak anisotropy in many clusters were also reported by Watkins
1644
+ et al. (2015a) and Watkins et al. (2015b), in which most of our
1645
+ samples were also studied. Watkins et al. (2015b) showed that their
1646
+ distance estimation had good agreement with Harris (1996) and con-
1647
+ cluded that the assumption of isotropy for their samples is reasonable.
1648
+ Watkins et al. (2015a) examined the ratio 𝜎T/𝜎R, which compared
1649
+ the tangential and radial components of the proper motion velocity
1650
+ dispersion at different radii. They found that the cluster centers are
1651
+ relatively isotropic, and the behavior of the increasing anisotropy
1652
+ with the radius was very moderate. From their figures, it can be seen
1653
+ that the decreasing of 𝜎T/𝜎R with a growing radius is more evident
1654
+ for NGC 5139 and NGC 7078.
1655
+ In recent years, Gaia has provided the proper motion data in the
1656
+ outer parts of globular clusters, and the behavior of 𝜎T/𝜎R reveals
1657
+ more evidence of anisotropy (Jindal et al. 2019; Vasiliev & Baum-
1658
+ gardt 2021). In both studies, NGC 5904 appears to be isotropic, and
1659
+ NGC 104, NGC 5139, and NGC 7078 show radial anisotropy. Some
1660
+ clusters are anisotropic in one study but are isotropic or uncertain
1661
+ in another; these include NGC 2808, NGC 6121, NGC 6397, NGC
1662
+ 6656, and NGC 6752.
1663
+ In addition, the anisotropy profiles 𝜎T/𝜎R−1 from the observations
1664
+ and our models are plotted in Fig. 7. The observational data was
1665
+ mainly from a recent report on the globular-cluster survey through
1666
+ Hubble Space Telescope (Libralato et al. 2022). It includes 16 clusters
1667
+ of our samples. The remaining two clusters were supplemented with
1668
+ the data from Watkins et al. (2015a). The data from Gaia (Jindal
1669
+ et al. 2019) which contains half of our samples were also used. In
1670
+ Fig. 7, the data of the above-discussed literature are expressed by
1671
+ open circles, crosses, and solid triangles; the profiles are roughly
1672
+ isotropic or slightly radial anisotropic within 𝑟 ≲ 100 arcsec. The
1673
+ larger anisotropy appears mainly in the outer regions. The radial
1674
+ anisotropy of NGC 5139 tends to increase from near 100 arcsec and
1675
+ later decrease to isotropy in 𝑟 ≳ 1000 arcsec. Our model predicts the
1676
+ decrease in radial anisotropy at a larger radius. For NGC 7078, the
1677
+ model shows a similar and milder profile to the observational one.
1678
+ NGC 6121 shows isotropy inside but grows to tangential anisotropy at
1679
+ a larger radius. The cluster was also found to be tangential anisotropy
1680
+ in Vasiliev & Baumgardt (2021). It could imply a more substantial
1681
+ influence from the tidal field, which is consistent with our results that
1682
+ this cluster has a smaller truncation parameter than others.
1683
+ 5.5 The Imprint of Galactic Tidal Field
1684
+ As mentioned earlier, the truncation parameter has the effect of mak-
1685
+ ing the extent of the system finite, and also drives the profile to be
1686
+ isotropic near the edge. These make the truncation parameter play
1687
+ a similar role as the external tidal field for the cluster. The exter-
1688
+ nal field generally becomes weaker for a larger distance from the
1689
+ Galactic center. Thus, clusters at larger distances from the Galactic
1690
+ center might be more extended and have larger values of truncation
1691
+ parameter 𝑔.
1692
+ In addition, Chernoff et al. (1986) found that the tidal field can
1693
+ increase the evolution rate of the cluster through relaxation and shock
1694
+ heating. Therefore, clusters closer to the Galactic center tend to evolve
1695
+ faster. They also suggested that inner regions of the Galaxy could be
1696
+ good places to look for the core-collapsed clusters. This agreed with
1697
+ Djorgovski & King (1986) who found that the mean and median
1698
+ distances of core-collapsed clusters from the Galactic center are
1699
+ smaller than 5 kpc.
1700
+ Moreover, the simulation in Zocchi et al. (2016) showed some
1701
+ related properties during the evolution of a globular cluster in an
1702
+ external tidal field. For example, the truncation parameter 𝑔 and the
1703
+ cluster mass 𝑀 decrease during the evolution. The concentration
1704
+ parameter 𝑊0 grows with time and decreases slightly after core col-
1705
+ lapse. The half-mass radius 𝑟h also increases with time and decreases
1706
+ as the cluster loses most of its mass.
1707
+ Motivated by the above results, here we examine possible correla-
1708
+ tions between any pairs among the concentration parameter 𝑊0, the
1709
+ truncation parameter 𝑔, the cluster mass 𝑀, the half-mass radius 𝑟h,
1710
+ and the semimajor axis of the cluster orbit 𝑎. The values of 𝑎 were
1711
+ taken as the average of the apogalactic and perigalactic distances in
1712
+ Baumgardt et al. (2019a), and the rest are our best-fit values in Table
1713
+ 1. The Spearman rank-order correlation coefficients, 𝐶s, were then
1714
+ calculated for all possible combinations; there were only two pairs
1715
+ with an absolute value of 𝐶s greater than 0.5. The first pair is the
1716
+ concentration parameter 𝑊0 and the truncation parameter 𝑔. Their
1717
+ 𝐶s = −0.65 indicates a strong anti-correlation between 𝑊0 and 𝑔.
1718
+ The distribution is presented in Fig. 8. The second pair is the trunca-
1719
+ tion parameter 𝑔 and the semimajor axis 𝑎 of the cluster orbit. The
1720
+ corresponding correlation coefficient 𝐶s = 0.60 indicates a strong
1721
+ correlation between 𝑔 and 𝑎; the result is presented in Fig. 9.
1722
+ The anti-correlation between the concentration parameter 𝑊0 and
1723
+ the truncation parameter 𝑔 is reasonable, as those with smaller trun-
1724
+ cation parameters would have experienced stronger tidal fields and
1725
+ evolve faster. It is likely that a certain fraction of them become
1726
+ core-collapsed clusters and thus have larger concentrations. This
1727
+ anti-correlation is also consistent with the simulations in Zocchi
1728
+ et al. (2016). They showed that when the clusters form, the value of
1729
+ concentration parameter 𝑊0 is nearly 4 and the value of truncation
1730
+ parameter 𝑔 is nearly 2.5. During the evolution, the truncation pa-
1731
+ rameter 𝑔 decreases, but the concentration parameter 𝑊0 increases.
1732
+ MNRAS 000, 1–15 (20XX)
1733
+
1734
+ Dynamical Properties of Globular Clusters
1735
+ 11
1736
+ 0.0
1737
+ 0.6
1738
+ 1.2
1739
+ 1.8
1740
+ 2.4
1741
+ g
1742
+ 10
1743
+ 0
1744
+ 10
1745
+ 20
1746
+ 30
1747
+ log ra
1748
+ 27.5
1749
+ 30.0
1750
+ 32.5
1751
+ 35.0
1752
+ M [105 M ]
1753
+ 7.5
1754
+ 9.0
1755
+ 10.5
1756
+ 12.0
1757
+ rh [pc]
1758
+ 5.16
1759
+ 5.22
1760
+ 5.28
1761
+ 5.34
1762
+ 5.40
1763
+ D [kpc]
1764
+ 0
1765
+ 10
1766
+ 20
1767
+ 30
1768
+ W0
1769
+ 1.8
1770
+ 2.4
1771
+ 3.0
1772
+ 3.6
1773
+ 4.2
1774
+ [
1775
+ ]
1776
+ 0.0
1777
+ 0.6
1778
+ 1.2
1779
+ 1.8
1780
+ 2.4
1781
+ g
1782
+ 10
1783
+ 0
1784
+ 10
1785
+ 20
1786
+ 30
1787
+ log ra
1788
+ 27.5
1789
+ 30.0
1790
+ 32.5
1791
+ 35.0
1792
+ M [105 M ]
1793
+ 7.5
1794
+ 9.0
1795
+ 10.5
1796
+ 12.0
1797
+ rh [pc]
1798
+ 5.16
1799
+ 5.22
1800
+ 5.28
1801
+ 5.34
1802
+ 5.40
1803
+ D [kpc]
1804
+ 1.8
1805
+ 2.4
1806
+ 3.0
1807
+ 3.6
1808
+ 4.2
1809
+ [
1810
+ ]
1811
+ Figure 5. The MCMC posterior parameter distributions of NGC 5139.
1812
+ Table 3. The parameters of two models of NGC 5139. The first column indicates different models. The second to eighth columns show the fitting parameters.
1813
+ The quantities in the last two columns are 𝜅 and 𝜒2r .
1814
+ Model
1815
+ 𝑊0
1816
+ 𝑔
1817
+ log ˆ𝑟a
1818
+ 𝑀
1819
+ 𝑟h
1820
+ 𝐷
1821
+ Υ
1822
+ 𝜅
1823
+ 𝜒2r
1824
+ (105 M⊙)
1825
+ (pc)
1826
+ (kpc)
1827
+ (Υ⊙)
1828
+ A
1829
+ 4.02+0.48
1830
+ −0.65
1831
+ 1.94+0.27
1832
+ −0.26
1833
+ 0.41+0.08
1834
+ −0.10
1835
+ 32.82+0.65
1836
+ −0.67
1837
+ 8.82+0.19
1838
+ −0.17
1839
+ 5.32 ± 0.03
1840
+ 2.38 ± 0.09
1841
+ 1.15
1842
+ 3.86
1843
+ B
1844
+ 14.16+0.25
1845
+ −0.23
1846
+ 1.28 ± 0.03
1847
+ 11.38+5.83
1848
+ −6.00
1849
+ 30.10 ± 0.59
1850
+ 10.260.17
1851
+ 0.16
1852
+ 5.25 ± 0.03
1853
+ 3.07 ± 0.09
1854
+ 1.00
1855
+ 5.64
1856
+ MNRAS 000, 1–15 (20XX)
1857
+
1858
+ 12
1859
+ Cheng and Jiang
1860
+ 101
1861
+ 102
1862
+ 103
1863
+ 15
1864
+ 20
1865
+ 25
1866
+ μ [mag/arcsec2]
1867
+ Model A
1868
+ 101
1869
+ 102
1870
+ 103
1871
+ 15
1872
+ 20
1873
+ 25
1874
+ Model B
1875
+ 101
1876
+ 102
1877
+ 103
1878
+ 0
1879
+ 10
1880
+ 20
1881
+ 30
1882
+ σlos [km/s]
1883
+ Model A
1884
+ 101
1885
+ 102
1886
+ 103
1887
+ 0
1888
+ 10
1889
+ 20
1890
+ 30
1891
+ Model B
1892
+ 101
1893
+ 102
1894
+ 103
1895
+ r [arcsec]
1896
+ 0
1897
+ 10
1898
+ 20
1899
+ 30
1900
+ σpm [km/s]
1901
+ Model A
1902
+ 101
1903
+ 102
1904
+ 103
1905
+ r [arcsec]
1906
+ 0
1907
+ 10
1908
+ 20
1909
+ 30
1910
+ Model B
1911
+ Figure 6. The comparison of the profiles from two models of NGC 5139. Left panels show the results from Model A and the right panels show those from
1912
+ Model B. The open circles represent the data of Baumgardt (2017), Baumgardt & Hilker (2018), and Dalgleish et al. (2020) for line-of-sight velocity dispersions
1913
+ and Watkins et al. (2015a) for proper motion velocity dispersions. The solid triangles are used for the data of Kamann et al. (2018) for line-of-sight velocity
1914
+ dispersions and Vasiliev & Baumgardt (2021) for proper motion velocity dispersions. The models are expressed by grey lines.
1915
+ Therefore, in Fig. 8, younger clusters are located at the top-left cor-
1916
+ ner, and the older clusters are distributed at the bottom-right corner.
1917
+ However, the exact relationships between these two parameters for
1918
+ different clusters are still complicated and the strength of this anti-
1919
+ correlation was not quantitatively investigated before.
1920
+ On the other hand, the correlation between the truncation parame-
1921
+ ter 𝑔 and the semimajor axis 𝑎 can be easily understood. The smaller
1922
+ truncation parameter shows that a stronger tidal field influences the
1923
+ cluster, and those clusters with smaller 𝑎 do experience stronger
1924
+ tidal fields. However, the relation between the two above-mentioned
1925
+ parameters shall also depends on the initial size and the orbital evolu-
1926
+ tion of a cluster. The contribution from different Galactic components
1927
+ make the exact behavior of the tidal field more complicated. It is rea-
1928
+ sonable that this correlation has a correlation coefficient 𝐶s = 0.60.
1929
+ The strong 𝑊0 − 𝑔 anti-correlation and 𝑔 − 𝑎 correlation shall
1930
+ be regarded as observational results as the employed parameters are
1931
+ obtained through our data-model fitting or from an observational
1932
+ catalog in literature. In addition, these observational anti-correlation
1933
+ and correlation agree with theoretical predictions.
1934
+ 6 SUMMARY AND CONCLUSIONS
1935
+ In this work, we studied 18 clusters with the LIMEPY models, a unified
1936
+ family of isothermal models. It can generate clusters with differ-
1937
+ ent amounts of concentration, truncation, and anisotropy, which are
1938
+ parametrized by continuous real numbers. Including some current
1939
+ observational data, such as the MUSE survey and Gaia mission, the
1940
+ fittings were carried out with a Markov Chain Monte Carlo ensemble
1941
+ sampler EMCEE and the parameters were determined by minimizing
1942
+ the 𝜒2 of the fittings.
1943
+ The measurable physical properties such as masses and distances,
1944
+ were compared with the values from the literature. Usually, Baum-
1945
+ gardt & Hilker (2018) has larger masses, while Watkins et al. (2015b)
1946
+ has smaller ones, and our results are in between. The smaller half-
1947
+ mass radius in our results is consistent with the smaller mass esti-
1948
+ mated compared with Baumgardt & Hilker (2018). Some differences
1949
+ between the radius estimations might come from the effect of the
1950
+ mass spectrum. For distance, our estimations are in agreement with
1951
+ the literature. The mass-to-light ratios are also similar to the litera-
1952
+ ture.
1953
+ Generally, the models could produce profiles similar to the obser-
1954
+ vational ones for most clusters. For NGC 5139, there are two groups
1955
+ of parameters that correspond to a better fitting for the surface bright-
1956
+ ness or the velocity-dispersion profiles. The anisotropic model gives a
1957
+ smaller 𝜒2r and agrees with the best-fit results in Zocchi et al. (2017).
1958
+ Some possible central dark objects, like an intermediate-mass black
1959
+ hole or a group of stellar-mass black holes might improve the fitting.
1960
+ NGC 6388 is also a candidate to host an intermediate-mass black
1961
+ hole, with the actual central line-of-sight velocities being uncertain.
1962
+ The data we used have the extension to nearly 5 arcsec with a velocity
1963
+ dispersion ∼20 km/s. It could be fitted well with the LIMEPY model
1964
+ except for the slope of proper-motion velocity dispersion.
1965
+ For the anisotropy, NGC 5139 and NGC 7078 are anisotropic
1966
+ MNRAS 000, 1–15 (20XX)
1967
+
1968
+ Dynamical Properties of Globular Clusters
1969
+ 13
1970
+ 101
1971
+ 102
1972
+ 103
1973
+ −1
1974
+ 0
1975
+ 1
1976
+ σT/σR − 1
1977
+ NGC 104
1978
+ 101
1979
+ 102
1980
+ −0.5
1981
+ 0.0
1982
+ 0.5
1983
+ NGC 288
1984
+ 101
1985
+ 102
1986
+ −0.5
1987
+ 0.0
1988
+ 0.5
1989
+ NGC 362
1990
+ 101
1991
+ 102
1992
+ −0.5
1993
+ 0.0
1994
+ 0.5
1995
+ σT/σR − 1
1996
+ NGC 1851
1997
+ 101
1998
+ 102
1999
+ −0.5
2000
+ 0.0
2001
+ 0.5
2002
+ NGC 2808
2003
+ 101
2004
+ 102
2005
+ −0.5
2006
+ 0.0
2007
+ 0.5
2008
+ NGC 3201
2009
+ 101
2010
+ 102
2011
+ 103
2012
+ −0.5
2013
+ 0.0
2014
+ 0.5
2015
+ σT/σR − 1
2016
+ NGC 5139
2017
+ 101
2018
+ 102
2019
+ 103
2020
+ −2
2021
+ −1
2022
+ 0
2023
+ 1
2024
+ 2
2025
+ NGC 5904
2026
+ 101
2027
+ 102
2028
+ 103
2029
+ −1
2030
+ 0
2031
+ 1
2032
+ NGC 6121
2033
+ 101
2034
+ 102
2035
+ −0.5
2036
+ 0.0
2037
+ 0.5
2038
+ σT/σR − 1
2039
+ NGC 6218
2040
+ 101
2041
+ 102
2042
+ −0.5
2043
+ 0.0
2044
+ 0.5
2045
+ NGC 6266
2046
+ 101
2047
+ 102
2048
+ −0.5
2049
+ 0.0
2050
+ 0.5
2051
+ NGC 6388
2052
+ 101
2053
+ 102
2054
+ 103
2055
+ −1
2056
+ 0
2057
+ 1
2058
+ σT/σR − 1
2059
+ NGC 6397
2060
+ 101
2061
+ 102
2062
+ −0.5
2063
+ 0.0
2064
+ 0.5
2065
+ NGC 6441
2066
+ 101
2067
+ 102
2068
+ 103
2069
+ −1
2070
+ 0
2071
+ 1
2072
+ NGC 6656
2073
+ 101
2074
+ 102
2075
+ r [arcsec]
2076
+ −0.5
2077
+ 0.0
2078
+ 0.5
2079
+ σT/σR − 1
2080
+ NGC 6715
2081
+ 101
2082
+ 102
2083
+ 103
2084
+ r [arcsec]
2085
+ −0.5
2086
+ 0.0
2087
+ 0.5
2088
+ NGC 6752
2089
+ 101
2090
+ 102
2091
+ r [arcsec]
2092
+ −1
2093
+ 0
2094
+ 1
2095
+ NGC 7078
2096
+ Figure 7. The anisotropy profiles of the clusters. The open circles represent the data of Libralato et al. (2022). The data of Jindal et al. (2019) are shown in
2097
+ solid triangles. The crosses are used for Watkins et al. (2015a). The models are expressed by grey lines. The horizontal grey dashed lines represent zeros which
2098
+ indicate isotropy. For each panel, the name of the cluster is mentioned at the top-left corner.
2099
+ MNRAS 000, 1–15 (20XX)
2100
+
2101
+ 14
2102
+ Cheng and Jiang
2103
+ 4
2104
+ 5
2105
+ 6
2106
+ 7
2107
+ 8
2108
+ 9
2109
+ W0
2110
+ 0.5
2111
+ 1.0
2112
+ 1.5
2113
+ 2.0
2114
+ 2.5
2115
+ g
2116
+ Figure 8. The truncation parameter versus the concentration parameter. The
2117
+ vertical axis represents the truncation parameter and the horizontal axis ex-
2118
+ presses the concentration parameter. Each point corresponds to a particular
2119
+ cluster.
2120
+ 5
2121
+ 10
2122
+ 15
2123
+ 20
2124
+ 25
2125
+ a [kpc]
2126
+ 0.5
2127
+ 1.0
2128
+ 1.5
2129
+ 2.0
2130
+ 2.5
2131
+ g
2132
+ Figure 9. The truncation parameter versus the semimajor axis of the cluster.
2133
+ The vertical axis represents the truncation parameter and the horizontal axis
2134
+ expresses the semimajor axis. Each data point corresponds to a particular
2135
+ cluster.
2136
+ with 𝜅 = 1.15 and 𝜅 = 1.16. The anisotropy leads to the rise in
2137
+ central velocity dispersion in these clusters. Our estimations could
2138
+ have some underestimations because the data are combined proper
2139
+ motion dispersion profiles rather than separated radial and tangential
2140
+ profiles. Nevertheless, the results are reasonable compared with some
2141
+ literature, such as Watkins et al. (2015a) and Watkins et al. (2015b),
2142
+ where the anisotropy in the studied clusters seem small.
2143
+ From a theoretical aspect, the truncation parameter may render
2144
+ the cluster to have a finite extension and isotropic profiles near the
2145
+ edge. It is similar to the effect of the external tidal field. In addition, a
2146
+ strong anti-correlation between the concentration parameter 𝑊0 and
2147
+ the truncation parameter 𝑔 was confirmed, which gives the imprint
2148
+ of the dynamical evolution of clusters. Finally, a strong correlation
2149
+ between the truncation parameter 𝑔 and the semimajor axis 𝑎 was
2150
+ also found, which could result from the influence of the Galactic tidal
2151
+ field.
2152
+ ACKNOWLEDGEMENTS
2153
+ We are grateful to the reviewer, Holger Baumgardt, for the useful sug-
2154
+ gestions which improved this paper significantly. We acknowledge
2155
+ the financial support from the Ministry of Science and Technology,
2156
+ Taiwan, (MOST grant 110-2112-M-007-035). We are grateful to the
2157
+ authors of Trager et al. (1995), Harris (1996), McLaughlin & van der
2158
+ Marel (2005), Baumgardt (2017), Baumgardt & Hilker (2018), Dal-
2159
+ gleish et al. (2020), Kamann et al. (2018), McLaughlin et al. (2006),
2160
+ Watkins et al. (2015a), Vasiliev & Baumgardt (2021), Häberle et al.
2161
+ (2021), McNamara et al. (2003), McNamara et al. (2012), Zloczewski
2162
+ et al. (2012), Watkins et al. (2015b), Baumgardt & Vasiliev (2021),
2163
+ Baumgardt et al. (2020), Libralato et al. (2022), Jindal et al. (2019),
2164
+ Baumgardt et al. (2019a), for making their data publicly available.
2165
+ This paper used the VizieR catalogue access tool, operated at CDS,
2166
+ Strasbourg, France, and the Astrophysics Data System Bibliographic
2167
+ Services of National Aeronautics Space and Administration, USA.
2168
+ Software: LIMEPY (Gieles & Zocchi 2015), EMCEE (Foreman-Mackey
2169
+ et al. 2013), corner, NumPy, and SciPy.
2170
+ DATA AVAILABILITY
2171
+ The electronic file of Table 1 is available in machine-readable form at
2172
+ VizieR (vizier.u-strasbg.fr) of Strasbourg astronomical Data Center
2173
+ (CDS).
2174
+ REFERENCES
2175
+ Bailyn C. D., 1995, ARA&A, 33, 133
2176
+ Baumgardt H., 2017, MNRAS, 464, 2174
2177
+ Baumgardt H., Hilker M., 2018, MNRAS, 478, 1520
2178
+ Baumgardt H., Makino J., 2003, MNRAS, 340, 227
2179
+ Baumgardt H., Vasiliev E., 2021, MNRAS, 505, 5957
2180
+ Baumgardt H., Makino J., Hut P., 2005, ApJ, 620, 238
2181
+ Baumgardt H., Hilker M., Sollima A., Bellini A., 2019a, MNRAS, 482, 5138
2182
+ Baumgardt H., et al., 2019b, MNRAS, 488, 5340
2183
+ Baumgardt H., Sollima A., Hilker M., 2020, Publ. Astron. Soc. Australia, 37,
2184
+ e046
2185
+ Bellazzini M., et al., 2008, AJ, 136, 1147
2186
+ Bianchini P., Ibata R., Famaey B., 2019, ApJ, 887, L12
2187
+ Blandford R., Meier D., Readhead A., 2019, ARA&A, 57, 467
2188
+ Chernoff D. F., Kochanek C. S., Shapiro S. L., 1986, ApJ, 309, 183
2189
+ Da Costa G. S., Freeman K. C., 1976, ApJ, 206, 128
2190
+ Dalgleish H., et al., 2020, MNRAS, 492, 3859
2191
+ Djorgovski S., King I. R., 1986, ApJ, 305, L61
2192
+ Ebisuzaki T., et al., 2001, ApJ, 562, L19
2193
+ Fitzpatrick E. L., 1999, PASP, 111, 63
2194
+ Foreman-Mackey D., Hogg D. W., Lang D., Goodman J., 2013, PASP, 125,
2195
+ 306
2196
+ GRAVITY Collaboration et al., 2019, A&A, 625, L10
2197
+ Gerssen J., van der Marel R. P., Gebhardt K., Guhathakurta P., Peterson R. C.,
2198
+ Pryor C., 2002, AJ, 124, 3270
2199
+ Gieles M., Zocchi A., 2015, MNRAS, 454, 576
2200
+ Gill M., Trenti M., Miller M. C., van der Marel R., Hamilton D., Stiavelli M.,
2201
+ 2008, ApJ, 686, 303
2202
+ Gomez-Leyton Y. J., Velazquez L., 2014, Journal of Statistical Mechanics:
2203
+ Theory and Experiment, 2014, 04006
2204
+ Goodman J., Weare J., 2010, Communications in Applied Mathematics and
2205
+ Computational Science, 5, 65
2206
+ MNRAS 000, 1–15 (20XX)
2207
+
2208
+ Dynamical Properties of Globular Clusters
2209
+ 15
2210
+ Göttgens F., et al., 2021, MNRAS, 507, 4788
2211
+ Gunn J. E., Griffin R. F., 1979, AJ, 84, 752
2212
+ Häberle M., et al., 2021, MNRAS, 503, 1490
2213
+ Harris W. E., 1996, AJ, 112, 1487
2214
+ Jindal A., Webb J. J., Bovy J., 2019, MNRAS, 487, 3693
2215
+ Kamann S., et al., 2018, MNRAS, 473, 5591
2216
+ King I. R., 1966, AJ, 71, 64
2217
+ Lanzoni B., et al., 2013, ApJ, 769, 107
2218
+ Libralato M., et al., 2022, ApJ, 934, 150
2219
+ Lützgendorf N., Kissler-Patig M., Noyola E., Jalali B., de Zeeuw P. T., Geb-
2220
+ hardt K., Baumgardt H., 2011, A&A, 533, A36
2221
+ Lynden-Bell D., 1967, MNRAS, 136, 101
2222
+ Manchester R. N., Lyne A. G., Robinson C., D’Amico N., Bailes M., Lim J.,
2223
+ 1991, Nature, 352, 219
2224
+ McLaughlin D. E., van der Marel R. P., 2005, ApJS, 161, 304
2225
+ McLaughlin D. E., Anderson J., Meylan G., Gebhardt K., Pryor C., Minniti
2226
+ D., Phinney S., 2006, ApJS, 166, 249
2227
+ McNamara B. J., Harrison T. E., Anderson J., 2003, ApJ, 595, 187
2228
+ McNamara B. J., Harrison T. E., Baumgardt H., Khalaj P., 2012, ApJ, 745,
2229
+ 175
2230
+ Michie R. W., 1963, MNRAS, 125, 127
2231
+ Mikolajewska J., Zdziarski A. A., Ziolkowski J., Torres M. A. P., Casares J.,
2232
+ 2022, ApJ, 930, 9
2233
+ Noyola E., Gebhardt K., Kissler-Patig M., Lützgendorf N., Jalali B., de Zeeuw
2234
+ P. T., Baumgardt H., 2010, ApJ, 719, L60
2235
+ Oh K. S., Lin D. N. C., 1992, ApJ, 386, 519
2236
+ Oort J. H., van Herk G., 1959, Bull. Astron. Inst. Netherlands, 14, 299
2237
+ Peuten M., Zocchi A., Gieles M., Gualandris A., Hénault-Brunet V., 2016,
2238
+ MNRAS, 462, 2333
2239
+ Peuten M., Zocchi A., Gieles M., Hénault-Brunet V., 2017, MNRAS, 470,
2240
+ 2736
2241
+ Plummer H. C., 1911, MNRAS, 71, 460
2242
+ Sanna N., Pancino E., Zocchi A., Ferraro F. R., Stetson P. B., 2020, A&A,
2243
+ 637, A46
2244
+ Spitzer Lyman J., Shapiro S. L., 1972, ApJ, 173, 529
2245
+ Takahashi K., Lee H. M., Inagaki S., 1997, MNRAS, 292, 331
2246
+ Tiongco M. A., Vesperini E., Varri A. L., 2016, MNRAS, 455, 3693
2247
+ Trager S. C., King I. R., Djorgovski S., 1995, AJ, 109, 218
2248
+ Vandenberg D. A., Bolte M., Stetson P. B., 1996, ARA&A, 34, 461
2249
+ Vasiliev E., Baumgardt H., 2021, MNRAS, 505, 5978
2250
+ Wan Z., et al., 2021, MNRAS, 502, 4513
2251
+ Watkins L. L., van der Marel R. P., Bellini A., Anderson J., 2015a, ApJ, 803,
2252
+ 29
2253
+ Watkins L. L., van der Marel R. P., Bellini A., Anderson J., 2015b, ApJ, 812,
2254
+ 149
2255
+ Wilson C. P., 1975, AJ, 80, 175
2256
+ Woolley R. V. D. R., 1954, MNRAS, 114, 191
2257
+ Zloczewski K., Kaluzny J., Rozyczka M., Krzeminski W., Mazur B., 2012,
2258
+ Acta Astron., 62, 357
2259
+ Zocchi A., Gieles M., Hénault-Brunet V., Varri A. L., 2016, MNRAS, 462,
2260
+ 696
2261
+ Zocchi A., Gieles M., Hénault-Brunet V., 2017, MNRAS, 468, 4429
2262
+ den Brok M., van de Ven G., van den Bosch R., Watkins L., 2014, MNRAS,
2263
+ 438, 487
2264
+ van Leeuwen F., 2009, A&A, 497, 209
2265
+ This paper has been typeset from a TEX/LATEX file prepared by the author.
2266
+ MNRAS 000, 1–15 (20XX)
2267
+
9tE4T4oBgHgl3EQfDQub/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
ANFKT4oBgHgl3EQfVi5k/content/tmp_files/2301.11788v1.pdf.txt ADDED
@@ -0,0 +1,679 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Prepared for submission to JHEP
2
+ Higgs Inflation: constraining the top quark mass
3
+ and breaking the H0-σ8 correlation
4
+ Jamerson G. Rodrigues,a Micol Benetti,b,c Rayff de Souzaa and Jailson Alcaniza
5
+ aObservatório Nacional, 20921-400, Rio de Janeiro, RJ, Brazil
6
+ bScuola Superiore Meridionale, Largo San Marcellino 10, 80138, Napoli, Italy
7
+ cIstituto Nazionale di Fisica Nucleare (INFN) Sezione di Napoli, Complesso Universitario di Monte
8
+ Sant’Angelo, Edificio G, Via Cinthia, I-80126, Napoli, Italy
9
10
11
+ Abstract: Extending previous results [JHEP 11 (2021) 091], we explore aspects of the
12
+ reheating mechanism for non-minimal Higgs inflation in the strong coupling regime. We
13
+ constrain the radiative corrections for the inflaton’s potential by considering the Coleman-
14
+ Weinberg approximation and use the Renormalization Group Equations for the Higgs field
15
+ to derive an upper limit on the quark top mass, mt. Using the current Cosmic Microwave
16
+ Background, Barion Acoustic Oscillation, and Supernova data, we obtain mt ≤ 170.44 GeV,
17
+ confirming the observational compatibility of the model with recent mt estimates reported
18
+ by the CMS collaboration. We also analyze the breakdown of the well-known correlation
19
+ involving the Hubble constant H0 and the clustering parameter σ8, which makes the model
20
+ interesting in light of the cosmological tensions discussed over the last decade.
21
+ Keywords: Cosmology, Primordial Universe, Cosmic Microwave Background, Higgs Field,
22
+ Cosmological Parameters.
23
+ arXiv:2301.11788v1 [astro-ph.CO] 27 Jan 2023
24
+
25
+ Contents
26
+ 1
27
+ Introduction
28
+ 1
29
+ 2
30
+ Non-minimal Inflation and Slow-Roll Analysis
31
+ 3
32
+ 3
33
+ Reheating analysis and results
34
+ 4
35
+ 4
36
+ Physical and cosmological consequences
37
+ 7
38
+ 4.1
39
+ Constraints on the top quark mass
40
+ 7
41
+ 4.2
42
+ The H0 − σ8 correlation
43
+ 7
44
+ 5
45
+ Conclusions
46
+ 9
47
+ 1
48
+ Introduction
49
+ The fundamental theory behind the initial conditions that led to the temperature fluctua-
50
+ tions in the Cosmic Microwave Background (CMB) [1, 2] and the formation of Large-Scale
51
+ Structure (LSS) of the universe [3–5] remains an open question in modern cosmology. In
52
+ this context, the paradigm of inflation rises as the most elegant description of the primordial
53
+ Universe [6–10]. In order to induce cosmic acceleration, the dynamical equations for the
54
+ inflaton field must enable a slowly varying solution, leading to a quasi-de Sitter Universe.
55
+ In the well-known slow-roll mechanism this is achieved in an approximately flat direction
56
+ of the inflaton’s scalar potential.
57
+ One particularly appealing approach is to induce a non-minimal coupling between the
58
+ inflaton and gravity, which results in a plateau at the large field regime [11–13] and drives
59
+ the model predictions to the sweet-spot of CMB observations [14]. From the phenomeno-
60
+ logical perspective, one specially interesting model was introduced by Berzrukov and Sha-
61
+ poshnikov [15], where the standard Higgs field rules the inflationary period at early times.
62
+ Such configuration allows one to compare the predictions of the model for the cosmological
63
+ observables with the phenomenology of the related particles at electroweak scale of energy.
64
+ Such analysis was explored in a number of interesting papers, see e.g. [16–20].
65
+ Although robust, the analysis of inflationary models rely on a set of assumptions about
66
+ the evolution of cosmological quantities. In particular, the evolution of cosmological scales
67
+ from the moment they cross the Hubble radius during inflation up to the their re-entrance
68
+ at later times must be matched to all the eras of the cosmological expansion in order to
69
+ solve the horizon problem [21]. The matching condition can be written in the form
70
+ ln
71
+
72
+ k
73
+ a0H0
74
+
75
+ = −Nk − Nrh − NRD + ln
76
+ �aeqHeq
77
+ a0H0
78
+
79
+ + ln
80
+ � Hk
81
+ Heq
82
+
83
+ ,
84
+ (1.1)
85
+ – 1 –
86
+
87
+ where Nk is the number of e-folds the universe expanded between the horizon crossing
88
+ moment of the pivot scale k and the end of inflation and Nrh is the number of e-folds counted
89
+ from the end of inflation to the onset of the radiation dominance in the early Universe
90
+ (reheating). Also, NRD gives the amount of expansion between the end of reheating and
91
+ the end of radiation dominated era, while the subscript “eq" and “0" represent quantities
92
+ evaluated at matter-radiation equality and the present, respectively. One is not able to set
93
+ the amount of expansion the universe experienced in the inflationary period, Nk, without
94
+ further information about the subsequent periods of the expansion. This is particularly
95
+ problematic for the reheating period.
96
+ In a previous communication [20], we performed a Monte-Carlo Markov Chain (MCMC)
97
+ analysis of CMB and clustering data to check the observational viability of non-minimally
98
+ coupled φ4 models for a fixed inflationary e-fold number. In particular, we considered the
99
+ first order correction to the perturbative expansion of the inflationary potential, also known
100
+ as Coleman-Weinberg approximation [22], and constrained possible radiative corrections
101
+ coming from the underlying field theory supporting this cosmological scenario. In addition,
102
+ we used the two-loop Renormalization Group Equations to connect the model’s predictions
103
+ at inflationary energy scales to the electroweak observables and derived an estimate of the
104
+ top quark mass mt, indicating a possible tension with the Monte-Carlo Tevatron and LHC
105
+ reconstruction [23].
106
+ In this work, we extend and complement the analysis reported in [20] by exploring the
107
+ predictions of non-minimal Higgs inflation for a wide range of the inflationary e-fold number
108
+ Nk and, consequently, of Nrh. Following the procedure developed in [20, 24], we employ a
109
+ MCMC analysis to compare the predictions of this inflationary scenario with the most recent
110
+ Cosmic Microwave Background (CMB), Baryon Acoustic Oscillation (BAO), and Supernova
111
+ (SN) data [1–5].
112
+ In particular, we obtain new constraints on the radiative corrections
113
+ coming from the underlying field theory supporting this cosmological scenario and derive
114
+ an upper limit for the top quark mass, which is compared with recent mt measurements
115
+ from different experiments. Furthermore, we also explore whether this model could shed
116
+ some light on the so-called cosmological tensions, which include the well-known H0 tension,
117
+ a ∼ 4σ-discrepancy between direct measurements of H0 using low-z SN (H0 = 73.48 ± 1.66
118
+ km/s/Mpc [25]) and the H0 estimate from current CMB data assuming the standard model
119
+ (H0 = 67.72 ± 0.41 km/s/Mpc [14]) [26, 27]. It is worth mentioning that most of the usual
120
+ mechanisms to solve this problem have failed so far, as alleviating the H0 discrepancy
121
+ worsens the agreement of other parameters with the data. In particular, the clustering
122
+ parameter, σ8, is constrained at σ8 = 0.766+0.024
123
+ −0.021 by the Kilo-Degree Survey (KiDS-1000)
124
+ lensing estimation [28] and its correlation with the Hubble constant leads to significantly
125
+ too high values as the value of H0 increases. Breaking such a correlation is not only tricky
126
+ but also challenging for many cosmological scenarios.
127
+ This work is organized as follows.
128
+ In Sec. 2, we briefly introduce the non-minimal
129
+ inflationary scenario and present the results of the slow-roll analysis. In Sec. 3, we discuss
130
+ aspects of the reheating stage following the Higgs inflation and present the main results of
131
+ our statistical analysis of the cosmological data. Sec. 4 discusses the constraints derived on
132
+ the top quark mass and some implications on the current cosmological tensions. The main
133
+ – 2 –
134
+
135
+ conclusions of this work are presented in Sec. 5.
136
+ 2
137
+ Non-minimal Inflation and Slow-Roll Analysis
138
+ As mentioned earlier, a common method to achieve slow-roll inflation is to induce a non-
139
+ minimal coupling between the inflaton field and gravity. Such procedure yields non-canonical
140
+ terms for the original scalar field and the metric, suggesting the use of a set of conformal
141
+ transformations in order to obtain the theory description in the familiar Einstein-Hilbert
142
+ formalism. A more detailed exposition of this approach can be found in [20].
143
+ The Einstein frame lagrangian reads
144
+ LE = −M2
145
+ P ˜R
146
+ 2
147
+ + 1
148
+ 2(∂µχ)†(∂µχ) − VE(χ) ,
149
+ (2.1)
150
+ and the subsequent time evolution is dictated by the inflaton’s potential
151
+ VE(χ) = λM4
152
+ P
153
+ 4ξ2
154
+
155
+ 1 − e−
156
+
157
+ 2
158
+ 3
159
+ χ
160
+ MP
161
+ �2
162
+
163
+ �1 + a′ ln
164
+
165
+ 1
166
+ ξ e
167
+
168
+ 2
169
+ 3
170
+ χ
171
+ MP − 1
172
+ ξ
173
+
174
+
175
+ (2.2)
176
+ where the large field regime is assumed for the inflaton, χ ≫
177
+
178
+ 6MP , and a large coupling
179
+ regime is assumed for the non-minimal coupling, ξ ≫ 1. Note that the deviation from
180
+ the tree level potential is quantified by the parameter a′ ≡ βλ/λ, where βλ is the running
181
+ equation for the quartic coupling λ. The above potential was obtained by adopting the
182
+ prescription II procedure to compute the radiative corrections in the Jordan frame and all
183
+ couplings are computed at the scale M = MP , where MP is the reduced Planck mass [20].
184
+ Once with the effective potential in the Einstein frame, the relevant slow-roll inflation-
185
+ ary parameters can be readily computed, which can be related to the spectral index and
186
+ tensor-to-scalar ratio, characteristic of the power spectrum of CMB perturbations probed
187
+ by Planck [1]. Although the field strength χ∗, necessary to compute the relevant inflation-
188
+ ary parameters, cannot be measured directly, we can infer its value from the duration of
189
+ inflation from horizon crossing up to the end of inflationary expansion, characterized by the
190
+ number of e-folds, which is also dependent on the form of the potential (2.2).
191
+ However, the inflationary number of e-folds is not a free parameter entirely, as it is
192
+ tied to the subsequent evolution of the universe, given its association with the horizon exit
193
+ of relevant cosmological scales. Therefore, the relevant scales probed by Planck seem to
194
+ correspond to an interval of 50-60 e-folds [21], which guides our range of exploration of the
195
+ parameter Nk.
196
+ In Fig. 1 we present our results for the spectral index and tensor-to-scalar ratio in the
197
+ nS × r plane, with a′ ranging from −0.1 (lower limit) to 1.0 (upper limit)1. Note that there
198
+ is a significant dependence of the inflationary predictions with the amount of expansion
199
+ during inflation, achieving compatibility with the Planck result2. It is also important to
200
+ 1The values of a′ varying between [-0.010, 0.053], [-0.020, 0.036] and [-0.027, 0.023], corresponding to
201
+ Nk = 50, 55 and 60, respectively, are in agreement with the 95% C.L. Planck result.
202
+ 2This agreement relies on the slow-roll approximations for the inflationary parameters and the phe-
203
+ nomenological power-law expansion of the primordial power spectrum.
204
+ – 3 –
205
+
206
+ Figure 1. ns vs. r for Nk = 50, 55 & 60. The points in each curve indicate the parameters for
207
+ a null resultant of the radiative corrections (a′ = 0). The blue areas show the favored regions by
208
+ Planck 2018, with 68% and 95% confidence level (Planck TT, TE, EE + lowE + lensing + BK15
209
+ + BAO data set) [14].
210
+ mention that the results obtained for the prediction of inflationary parameters are highly
211
+ independent of the coupling parameter ξ.
212
+ 3
213
+ Reheating analysis and results
214
+ Between the end of inflation and the onset of a radiation-dominated universe, the universe
215
+ undergoes a reheating period. Even though there are a number of proposals for the dy-
216
+ namics of the cosmos in this period [29–36], the reheating era is exceptionally difficult to
217
+ be constrained by observations, given the small length scales characteristic of this micro-
218
+ physical process. For previous works exploring the impact of reheating to the cosmological
219
+ observables see e.g. [37–40] and references therein.
220
+ In order to understand the influence of the reheating period on the inflationary predic-
221
+ tions, one can follow the steps developed in [38] and resume the matching condition (1.1)
222
+ to the expression:
223
+ Nk = −1 + 3ωrh
224
+ 4
225
+ Nrh − ln
226
+
227
+ V 1/4
228
+ end
229
+ Hk
230
+
231
+ + 61.55 ,
232
+ (3.1)
233
+ where the amount of expansion through the inflationary period is explicitly related to the
234
+ reheating characteristics of the proposed model. Here, ωrh represents the effective equation-
235
+ of-state parameter of the cosmological fluid during reheating, Vend is the amplitude of the
236
+ inflaton’s potential energy at the end of inflation, Hk is the Hubble parameter evaluated at
237
+ horizon crossing and k = 0.05 Mpc−1 is the pivot scale. We also consider grh ∼ 100 for the
238
+ relativistic degrees of freedom to obtain the numerical factor above.
239
+ – 4 –
240
+
241
+ 0.08
242
+ Nx = 50
243
+ Nx = 55
244
+ 0.06
245
+ Nx = 60
246
+ 0.05
247
+ 0.03
248
+ 0.D2
249
+ 0.D1
250
+ 0.00
251
+ 0.95
252
+ 0.96
253
+ L60
254
+ 0.98
255
+ 0.99
256
+ fha′
257
+ r0.02
258
+ H0
259
+ σ8
260
+ Nk=50
261
+ 0.179 ± 0.072
262
+ 0.032 ± 0.013
263
+ 68.82 ± 0.38
264
+ 0.841 ± 0.005
265
+ Nk=52
266
+ 0.040 ± 0.015
267
+ 0.007 ± 0.002
268
+ 68.31 ± 0.41
269
+ 0.835 ± 0.005
270
+ Nk=54
271
+ 0.011 ± 0.014
272
+ 0.004 ± 0.001
273
+ 67.71 ± 0.45
274
+ 0.817 ± 0.003
275
+ Nk=54.5
276
+ 0.009 ± 0.013
277
+ 0.004 ± 0.001
278
+ 67.68 ± 0.43
279
+ 0.811 ± 0.003
280
+ Nk=55
281
+ 0.010 ± 0.013
282
+ 0.004 ± 0.001
283
+ 67.71 ± 0.44
284
+ 0.804 ± 0.003
285
+ Nk=56
286
+ 0.022 ± 0.015
287
+ 0.005 ± 0.001
288
+ 67.94 ± 0.45
289
+ 0.793 ± 0.003
290
+ Nk=58
291
+ 0.283 ± 0.169
292
+ 0.044 ± 0.019
293
+ 68.37 ± 0.39
294
+ 0.779 ± 0.004
295
+ Nk=60
296
+ 0.243 ± 0.088
297
+ 0.042 ± 0.015
298
+ 68.46 ± 0.38
299
+ 0.766 ± 0.005
300
+ Table 1. Constraints for fixed Nk at 68% C.L. using the Planck TT, TE, EE + lowE + lensing +
301
+ BICEP2/Keck + BAO + Pantheon combination.
302
+ In what concerns non-minimal inflationary models, it is possible to show that the
303
+ inflaton condensate starts the reheating process oscillating with an effective matter-like
304
+ equation of state (ω1 = 0) and, after crossing a critical value χcr, finishes the process as a
305
+ radiation-like component of energy (ω2 = 1/3) [41, 42]. After some algebraic manipulations
306
+ and using the approximation Hk ∼
307
+
308
+ V∗/3, valid during inflation, one obtains:
309
+ Nk = −1
310
+ 4N1 − ln
311
+
312
+ V 1/4
313
+ end (a′)
314
+
315
+ V∗(a′)/3
316
+
317
+ + 61.55
318
+ (3.2)
319
+ where we highlight the a′ dependence of the inflationary potential.
320
+ We analyze the present model for fixed values of Nk and compute the values of Vend
321
+ and Hk following the slow-roll approximations.
322
+ In our analysis we assume a standard
323
+ cosmological model with a modified primordial spectrum in which the radiative correction
324
+ parameter, a′, is free to vary.
325
+ For the parameter estimation we use the free available
326
+ CosmoMC code [43]3 and a combination of early and late data4 (for more details we refer
327
+ the reader to [20]). Table 1 shows the derived constraints on the most significant parameters
328
+ of our analysis.
329
+ Note that by computing the values of Vend and Hk, we can obtain the corresponding
330
+ values for N1, i.e, the amount of expansion that the universe went through, as matter-
331
+ like dominated, during the reheating process. The corresponding values are presented in
332
+ Figure 2. Note also that, for an expansion of ∼ 56 e-folds or greater during inflation, N1
333
+ would have to assume negative values to satisfy the matching equation (3.2). By definition,
334
+ this condition would imply in a contraction of the universe between the end of inflation
335
+ 3This is a MCMC code interfaced with the Boltzmann solver Code for Anisotropies in the Microwave
336
+ Background (CAMB) [44]. We modified CAMB following the indications of ModeCode [45, 46] in order to
337
+ analyse the specific form of the potential V (φ).
338
+ 4We use the CMB Planck (2018) likelihood [1], using Plik temperature power spectrum, TT, and HFI
339
+ polarization EE likelihood at ℓ ≤ 29; BICEP2 and Keck Array experiments B-mode polarization data [2];
340
+ BAO measurements from 6dFGS
341
+ [3], SDSS-MGS [47], and BOSS DR12 [4] surveys, and the Pantheon
342
+ sample of Type Ia supernovae [5].
343
+ – 5 –
344
+
345
+ Figure 2. Nk vs. N1 for each inflationary number of e-folds taken into consideration. N1 is given
346
+ by the matching equation (3.2), with a′ coming from the MCMC analysis (highlighted beside each
347
+ point). Through a linear regression between the points (solid blue line), we estimate a maximum
348
+ number Nk - where the transition to a radiation-dominated Universe happens instantaneously.
349
+ and the onset of the radiation-dominated epoch5. Thus, following the standard approach,
350
+ we discard these possibilities as non-physical. Therefore, we can tighten the bounds on
351
+ the maximum value for the inflationary number of e-folds, which yields an instantaneous
352
+ transition to the radiation-dominated expansion.
353
+ The results presented above are insensitive to the specific physical process that leads to
354
+ the transition between matter and radiation-like expansion in the reheating. As pointed out
355
+ in [17, 41], non-perturbative processes may occur before the perturbative decays become
356
+ viable (preheating), displacing the transition between the two expansion behaviors, which
357
+ is particularly true in the model of Higgs Inflation. In this context, a specially interesting
358
+ result was obtained in [48], where the authors discussed the resonant production of Higgs
359
+ and gauge degrees of freedom in the linear regime of the Higgs Inflation scenario.
360
+ For
361
+ 100 < ξ < 1000, the preheating dominant process is the Higgs self-resonance, leading to
362
+ N1 ≃ 3. For higher values of the non-minimal coupling, ξ > 1000, it was pointed out that a
363
+ substantial amount of energy stored in the inflaton condensate is transferred to relativistic
364
+ gauge bosons already at the very first oscillation of the background (instant preheating),
365
+ leading to N1 = 0. Note that these results are in agreement with our analysis for Nk ≃ 55
366
+ and Nk ≃ 56, respectively, which is also in agreeement with the MCMC result for the
367
+ 5It is also possible to obtain N1 > 0 even for Nk > 56 if one considers exotic scenarios for the transition to
368
+ radiation dominance, including intermediary phase transitions of the reheating fluid to an exotic component
369
+ of energy ω′ > 1/3.
370
+ – 6 –
371
+
372
+ 0.179
373
+ 20
374
+ *: Nk.max~55.98
375
+ 10.04
376
+ 10
377
+ M
378
+ N0.011
379
+ 10.009
380
+ 10.01
381
+ 0
382
+ 10.022
383
+ 0.283
384
+ -10
385
+ 0.243
386
+ 50
387
+ 52
388
+ 54
389
+ 55
390
+ 56
391
+ 58
392
+ 60
393
+ Nkradiative corrections in the interval a′ ≃ [−0.003, 0.037] at 68% (C.L.).
394
+ 4
395
+ Physical and cosmological consequences
396
+ 4.1
397
+ Constraints on the top quark mass
398
+ It is helpful to recall that the result mentioned above is obtained in the framework of the
399
+ Higgs Inflation scenario, where a′ is associated with the β-function of the Higgs quartic
400
+ coupling λ. Once the renormalization group equations for the standard Higgs couplings
401
+ are considered, it is possible to link the cosmological constraints to the phenomenology
402
+ of the associated particles at the electroweak scale of energy6. In this context, following
403
+ the approach developed in [20], one shall infer an upper limit on the top quark pole mass,
404
+ mt ≤ 170.44 GeV, to reproduce the values of a′ above. Also, it is worth emphasizing that
405
+ this limit on mt is relatively insensitive to the amplitude of the non-minimal coupling once
406
+ the strong limit (ξ ≫ 1) is assumed.
407
+ The most precise constraints on the top quark mass are extracted from the kinematic
408
+ reconstruction of the t¯t events where mt is employed in the Monte-Carlo generator in order
409
+ to fit the data [49, 50]. This MC top quark mass is usually assumed to be the pole mass
410
+ even though the theoretical uncertainties inherent to this association are hard to quantify
411
+ [51]. From [52], the average value for the top quark mass is set to mt = 172.69 ± 0.30 GeV,
412
+ obtained from LHC and Tevatron data. If contrasted with the limit on mt obtained from
413
+ the cosmological analysis, this represents a significant discrepancy of 7.5σ.
414
+ Instead, one may consider theoretically cleaner the inference of the top quark pole
415
+ mass from the measurements of the cross-section of the top quark production, since the
416
+ theoretical computation of σ(t¯t) is explicitly performed in a renormalization scheme (e.g.,
417
+ MS) [53]. In this case, the average value obtained from the Tevatron and LHC runs is
418
+ 172.5 ± 0.7 GeV [52], lowering the discrepancy with our cosmological estimate of mt to
419
+ ≈ 3σ. More recently, the CMS collaboration reported mt = 170.5±0.8 GeV, obtained from
420
+ the differential cross-section of the top production [54]. Such result perfectly agrees with
421
+ the results of our cosmological analysis of the Higgs Inflation.
422
+ 4.2
423
+ The H0 − σ8 correlation
424
+ The accuracy of cosmological and astrophysical measurements has significantly improved
425
+ in recent decades. While this has led to increasingly evident confirmation of the validity of
426
+ the standard cosmological model, it has also exposed some critical issues that have given
427
+ rise to heated debate. The well-known H0 tension has been extensively explored without
428
+ concluding so far (we refer the reader to [26, 27] and references therein).
429
+ It has also been widely pointed out that some of the current attempts to solve the
430
+ H0 tension have failed because as they alleviate the discrepancy on H0, they worsen the
431
+ agreement of other parameters with the data. In particular, the clustering parameter, σ8, is
432
+ constrained at σ8 = 0.766+0.024
433
+ −0.021 by the Kilo-Degree Survey (KiDS-1000) lensing estimation
434
+ 6The parameters considered in the definition of a′ are evaluated at the renormalization scale M = MP .
435
+ – 7 –
436
+
437
+ 0.75
438
+ 0.80
439
+ 0.85
440
+ 8
441
+ 66
442
+ 67
443
+ 68
444
+ 69
445
+ 70
446
+ H0
447
+ 0.75
448
+ 0.8
449
+ 0.85
450
+ 8
451
+ N=50
452
+ N=52
453
+ N=54
454
+ N=54.5
455
+ N=55
456
+ N=56
457
+ N=58
458
+ N=60
459
+ Figure 3. Confidence levels and posterior distributions for the H0 and σ8 parameters using the
460
+ joint data set CMB Planck (2018) + BICEP2 and Keck Array + BAO + Pantheon SNe Ia sample
461
+ and considering several values of Nk.
462
+ [28] and its correlation with the Hubble constant leads to values that are significantly too
463
+ high as the value of H0 increases.
464
+ It is generally agreed that a model that manages to resolve both tensions is a model
465
+ that breaks this degeneracy, but building such a model is proving difficult. So far, only
466
+ a handful of scenarios seem to succeed, such as the conjecture of a universe transition
467
+ from anti-de Sitter vacua to de Sitter vacua [55–57], some early interacting models [58] or
468
+ specific parametrizations of dark energy equation of state [59]. The model studied here is
469
+ promising, as it breaks the degeneracy between the two parameters. In particular, Table 1
470
+ shows that as Nk increases between the values of 50 and 54.5, the values of the radiative
471
+ parameter, a′, H0, and σ8 decrease. Nevertheless, at the turning value of Nk = 54.5, there
472
+ is a behavior change, i.e., as Nk increases, the values of a′ and H0 also increase. In contrast,
473
+ the value of the clustering parameter, σ8, does not seem to be affected by this turning point
474
+ and continues to decrease. It means that, for values of Nk ∈ [54.5, 60]7, the correlation
475
+ between H0 and σ8 breaks down, as also shown in Figure 3. In particular, for the limiting
476
+ value Nk = 56, i.e., an instantaneous transition to the radiation-dominated expansion, the
477
+ degeneracy H0 − σ8 is such that it reduces the H0 tension, constraining H0 = 67.94 ± 0.45
478
+ 7As discussed earlier, we consider the cases Nk > 56 to be non-physical since they predict negative values
479
+ of N1.
480
+ – 8 –
481
+
482
+ Km/s/Mpc, which is ≈ 3σ off from the SNe Ia measurements [25] and allowing a value of
483
+ σ8 = 0.793 ± 0.003, that is in full agreement with KiDS-1000 results [28].
484
+ 5
485
+ Conclusions
486
+ In this work, we revisited the non-minimal inflationary scenario subject to radiative cor-
487
+ rections.
488
+ By performing an observational analysis of the φ4 primordial potential, non-
489
+ minimally coupled to the Ricci scalar, in light of the most recent CMB, clustering and
490
+ Supernova data and considering the allowed range for the observable inflationary e-folds,
491
+ we constrained the possible values of the radiative corrections of the inflaton potential,
492
+ encoded in the parameter a′, and the usual set of cosmological parameters.
493
+ From this analysis, we presented two main results. First, we set an upper limit to the
494
+ number of e-folds from the horizon crossing moment up to the end of inflation, Nk ≲ 56,
495
+ relative to instantaneous reheating, by considering the matching equation for the pivot scale
496
+ k = 0.05 Mpc−1. An even more stringent limit is imposed once considered the preheating
497
+ structure of the Higgs Inflation, yielding 55 ≲ Nk ≲ 56. Accordingly, the MCMC analysis
498
+ of the model translates into an upper bound for the top quark pole mass, mt ≤ 170.44 GeV,
499
+ which raises two possible interpretations for the consistency of the model at low-energies.
500
+ For example, considering the value of the top quark mass reconstructed from the analysis of
501
+ LHC and Tevatron data, Mt = 172.69 ± 0.30 GeV [23], implies a significant tension of 7.5σ
502
+ between the observed low-energy value and the amount inferred by the cosmological MCMC
503
+ analysis. On the other hand, assuming the top quark mass extracted from differential cross-
504
+ section of the top production, Mt = 170.5 ± 0.8 GeV, obtained by the CMS collaboration
505
+ [54], we found a perfect agreement between the cosmological analysis of the Higgs field and
506
+ its electroweak behaviour.
507
+ Second, the MCMC analysis of current observational data confirms the observational
508
+ viability of the model and shows that for the interval Nk ∈ [54.5, 60], it can break down the
509
+ well-known H0 − σ8 correlation (see Table 1). In particular, considering an instantaneous
510
+ transition to the radiation-dominated expansion, which occurs for Nk = 56, the H0 tension
511
+ is reduced to ≈ 3σ whereas the value of σ8 shows a complete agreement with KiDS-1000
512
+ results.
513
+ These results reinforce the need to investigate Higgs inflation and its extensions from
514
+ both theoretical and observational sides and show that perspectives for a complete coherence
515
+ of the scenario may converge once data from future collider experiments [60, 61] improve
516
+ our understanding of the physics at the eletroweak scale.
517
+ Acknowledgments
518
+ We thank André Sznajder for helpful conversations. JGR acknowledges financial support
519
+ from the Programa de Capacitação Institucional (PCI) do Observatório Nacional/MCTI.
520
+ MB acknowledges Istituto Nazionale di Fisica Nucleare (INFN), sezione di Napoli, iniziativa
521
+ specifica QGSKY. RdS is supported by the Coordenação de Aperfeiçoamento de Pessoal
522
+ de Nível Superior (CAPES). JSA is supported by CNPq (Grants no. 310790/2014-0 and
523
+ – 9 –
524
+
525
+ 400471/2014-0) and Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro FAPERJ
526
+ (grant no. 233906). We also acknowledge the use of CosmoMC and ModeCode packages.
527
+ This work was developed thanks to the use of the National Observatory Data Center (CP-
528
+ DON).
529
+ References
530
+ [1] Planck Collaboration, N. Aghanim et al., Planck 2018 results. V. CMB power spectra and
531
+ likelihoods, Astron. Astrophys. 641 (2020) A5, [arXiv:1907.12875].
532
+ [2] BICEP2, Keck Array Collaboration, P. A. R. Ade et al., BICEP2 / Keck Array x:
533
+ Constraints on Primordial Gravitational Waves using Planck, WMAP, and New
534
+ BICEP2/Keck Observations through the 2015 Season, Phys. Rev. Lett. 121 (2018) 221301,
535
+ [[1810.05216]].
536
+ [3] F. Beutler, C. Blake, M. Colless, D. H. Jones, L. Staveley-Smith, L. Campbell, Q. Parker,
537
+ W. Saunders, and F. Watson, The 6dF Galaxy Survey: Baryon Acoustic Oscillations and the
538
+ Local Hubble Constant, Mon. Not. Roy. Astron. Soc. 416 (2011) 3017–3032,
539
+ [arXiv:1106.3366].
540
+ [4] BOSS Collaboration, S. Alam et al., The clustering of galaxies in the completed SDSS-III
541
+ Baryon Oscillation Spectroscopic Survey: cosmological analysis of the DR12 galaxy sample,
542
+ Mon. Not. Roy. Astron. Soc. 470 (2017), no. 3 2617–2652, [arXiv:1607.03155].
543
+ [5] D. M. Scolnic et al., The Complete Light-curve Sample of Spectroscopically Confirmed SNe Ia
544
+ from Pan-STARRS1 and Cosmological Constraints from the Combined Pantheon Sample,
545
+ Astrophys. J. 859 (2018), no. 2 101, [arXiv:1710.00845].
546
+ [6] A. H. Guth, The Inflationary Universe: A Possible Solution to the Horizon and Flatness
547
+ Problems, Phys. Rev. D23 (1981) 347–356.
548
+ [7] A. D. Linde, A New Inflationary Universe Scenario: A Possible Solution of the Horizon,
549
+ Flatness, Homogeneity, Isotropy and Primordial Monopole Problems, Phys. Lett. 108B
550
+ (1982) 389–393.
551
+ [8] A. Albrecht and P. J. Steinhardt, Cosmology for Grand Unified Theories with Radiatively
552
+ Induced Symmetry Breaking, Phys. Rev. Lett. 48 (1982) 1220–1223.
553
+ [9] V. F. Mukhanov and G. V. Chibisov, Quantum Fluctuations and a Nonsingular Universe,
554
+ JETP Lett. 33 (1981) 532–535.
555
+ [10] A. A. Starobinsky, Dynamics of Phase Transition in the New Inflationary Universe Scenario
556
+ and Generation of Perturbations, Phys. Lett. B 117 (1982) 175–178.
557
+ [11] A. A. Starobinsky, A New Type of Isotropic Cosmological Models Without Singularity, Phys.
558
+ Lett. 91B (1980) 99–102. [,771(1980)].
559
+ [12] B. L. Spokoiny, INFLATION AND GENERATION OF PERTURBATIONS IN BROKEN
560
+ SYMMETRIC THEORY OF GRAVITY, Phys. Lett. B 147 (1984) 39–43.
561
+ [13] D. S. Salopek, J. R. Bond, and J. M. Bardeen, Designing Density Fluctuation Spectra in
562
+ Inflation, Phys. Rev. D40 (1989) 1753.
563
+ [14] Planck Collaboration, N. Aghanim et al., Planck 2018 results. VI. Cosmological parameters,
564
+ Astron. Astrophys. 641 (2020) A6, [arXiv:1807.06209]. [Erratum: Astron.Astrophys. 652,
565
+ C4 (2021)].
566
+ – 10 –
567
+
568
+ [15] F. L. Bezrukov and M. Shaposhnikov, The Standard Model Higgs boson as the inflaton, Phys.
569
+ Lett. B659 (2008) 703–706, [arXiv:0710.3755].
570
+ [16] A. O. Barvinsky, A. Yu. Kamenshchik, and A. A. Starobinsky, Inflation scenario via the
571
+ Standard Model Higgs boson and LHC, JCAP 0811 (2008) 021, [arXiv:0809.2104].
572
+ [17] J. Garcia-Bellido, D. G. Figueroa, and J. Rubio, Preheating in the Standard Model with the
573
+ Higgs-Inflaton coupled to gravity, Phys. Rev. D79 (2009) 063531, [arXiv:0812.4624].
574
+ [18] F. Bezrukov and M. Shaposhnikov, Higgs inflation at the critical point, Phys. Lett. B734
575
+ (2014) 249–254, [arXiv:1403.6078].
576
+ [19] H. M. Lee, Light inflaton completing Higgs inflation, Phys. Rev. D98 (2018), no. 1 015020,
577
+ [arXiv:1802.06174].
578
+ [20] J. G. Rodrigues, M. Benetti, and J. S. Alcaniz, Possible discrepancies between cosmological
579
+ and electroweak observables in Higgs Inflation, JHEP 11 (2021) 091, [arXiv:2105.07009].
580
+ [21] A. R. Liddle and S. M. Leach, How long before the end of inflation were observable
581
+ perturbations produced?, Phys. Rev. D68 (2003) 103503, [astro-ph/0305263].
582
+ [22] S. R. Coleman and E. J. Weinberg, Radiative Corrections as the Origin of Spontaneous
583
+ Symmetry Breaking, Phys. Rev. D7 (1973) 1888–1910.
584
+ [23] Particle Data Group Collaboration, P. Zyla et al., Review of Particle Physics, PTEP
585
+ 2020 (2020), no. 8 083C01.
586
+ [24] J. G. Rodrigues, M. Benetti, M. Campista, and J. Alcaniz, Probing the Seesaw Mechanism
587
+ with Cosmological data, JCAP 07 (2020) 007, [arXiv:2002.05154].
588
+ [25] A. G. Riess et al., New Parallaxes of Galactic Cepheids from Spatially Scanning the Hubble
589
+ Space Telescope: Implications for the Hubble Constant, Astrophys. J. 855 (2018), no. 2 136,
590
+ [arXiv:1801.01120].
591
+ [26] E. Di Valentino, O. Mena, S. Pan, L. Visinelli, W. Yang, A. Melchiorri, D. F. Mota, A. G.
592
+ Riess, and J. Silk, In the realm of the Hubble tension—a review of solutions, Class. Quant.
593
+ Grav. 38 (2021), no. 15 153001, [arXiv:2103.01183].
594
+ [27] E. Di Valentino et al., Snowmass2021 - Letter of interest cosmology intertwined II: The
595
+ hubble constant tension, Astropart. Phys. 131 (2021) 102605, [arXiv:2008.11284].
596
+ [28] KiDS Collaboration, M. Asgari et al., KiDS-1000 Cosmology: Cosmic shear constraints and
597
+ comparison between two point statistics, Astron. Astrophys. 645 (2021) A104,
598
+ [arXiv:2007.15633].
599
+ [29] L. F. Abbott, E. Farhi, and M. B. Wise, Particle Production in the New Inflationary
600
+ Cosmology, Phys. Lett. 117B (1982) 29.
601
+ [30] A. D. Dolgov and A. D. Linde, Baryon Asymmetry in Inflationary Universe, Phys. Lett. B
602
+ 116 (1982) 329.
603
+ [31] A. Albrecht, P. J. Steinhardt, M. S. Turner, and F. Wilczek, Reheating an Inflationary
604
+ Universe, Phys. Rev. Lett. 48 (1982) 1437.
605
+ [32] L. Kofman, A. D. Linde, and A. A. Starobinsky, Reheating after inflation, Phys. Rev. Lett.
606
+ 73 (1994) 3195–3198, [hep-th/9405187].
607
+ [33] J. H. Traschen and R. H. Brandenberger, Particle Production During Out-of-equilibrium
608
+ Phase Transitions, Phys. Rev. D 42 (1990) 2491–2504.
609
+ – 11 –
610
+
611
+ [34] L. Kofman, A. D. Linde, and A. A. Starobinsky, Towards the theory of reheating after
612
+ inflation, Phys. Rev. D 56 (1997) 3258–3295, [hep-ph/9704452].
613
+ [35] Y. Shtanov, J. H. Traschen, and R. H. Brandenberger, Universe reheating after inflation,
614
+ Phys. Rev. D 51 (1995) 5438–5455, [hep-ph/9407247].
615
+ [36] M. A. Amin, M. P. Hertzberg, D. I. Kaiser, and J. Karouby, Nonperturbative Dynamics Of
616
+ Reheating After Inflation: A Review, Int. J. Mod. Phys. D 24 (2014) 1530003,
617
+ [arXiv:1410.3808].
618
+ [37] L. Dai, M. Kamionkowski, and J. Wang, Reheating constraints to inflationary models, Phys.
619
+ Rev. Lett. 113 (2014) 041302, [arXiv:1404.6704].
620
+ [38] J. L. Cook, E. Dimastrogiovanni, D. A. Easson, and L. M. Krauss, Reheating predictions in
621
+ single field inflation, JCAP 04 (2015) 047, [arXiv:1502.04673].
622
+ [39] M. Drewes, What can the CMB tell about the microphysics of cosmic reheating?, JCAP 03
623
+ (2016) 013, [arXiv:1511.03280].
624
+ [40] S. S. Mishra, V. Sahni, and A. A. Starobinsky, Curing inflationary degeneracies using
625
+ reheating predictions and relic gravitational waves, JCAP 05 (2021) 075,
626
+ [arXiv:2101.00271].
627
+ [41] F. Bezrukov, D. Gorbunov, and M. Shaposhnikov, On initial conditions for the Hot Big
628
+ Bang, JCAP 0906 (2009) 029, [arXiv:0812.3622].
629
+ [42] J.-O. Gong, S. Pi, and G. Leung, Probing reheating with primordial spectrum, JCAP 05
630
+ (2015) 027, [arXiv:1501.03604].
631
+ [43] A. Lewis and S. Bridle, Cosmological parameters from CMB and other data: A Monte Carlo
632
+ approach, Phys. Rev. D66 (2002) 103511, [astro-ph/0205436].
633
+ [44] A. Lewis, A. Challinor, and A. Lasenby, Efficient computation of CMB anisotropies in closed
634
+ FRW models, Astrophys. J. 538 (2000) 473–476, [astro-ph/9911177].
635
+ [45] M. J. Mortonson, H. V. Peiris, and R. Easther, Bayesian Analysis of Inflation: Parameter
636
+ Estimation for Single Field Models, Phys. Rev. D83 (2011) 043505, [arXiv:1007.4205].
637
+ [46] R. Easther and H. V. Peiris, Bayesian Analysis of Inflation II: Model Selection and
638
+ Constraints on Reheating, Phys. Rev. D85 (2012) 103533, [arXiv:1112.0326].
639
+ [47] A. J. Ross, L. Samushia, C. Howlett, W. J. Percival, A. Burden, and M. Manera, The
640
+ clustering of the SDSS DR7 main Galaxy sample I : A 4 per cent distance measure at
641
+ z = 0.15, Mon. Not. Roy. Astron. Soc. 449 (2015), no. 1 835–847, [arXiv:1409.3242].
642
+ [48] E. I. Sfakianakis and J. van de Vis, Preheating after Higgs Inflation: Self-Resonance and
643
+ Gauge boson production, Phys. Rev. D 99 (2019), no. 8 083519, [arXiv:1810.01304].
644
+ [49] CDF Collaboration, F. Abe et al., Evidence for top quark production in ¯pp collisions at
645
+ √s = 1.8 TeV, Phys. Rev. D 50 (1994) 2966–3026.
646
+ [50] D0 Collaboration, V. M. Abazov et al., A precision measurement of the mass of the top
647
+ quark, Nature 429 (2004) 638–642, [hep-ex/0406031].
648
+ [51] F. Bezrukov and M. Shaposhnikov, Why should we care about the top quark Yukawa
649
+ coupling?, J. Exp. Theor. Phys. 120 (2015) 335–343, [arXiv:1411.1923].
650
+ [52] Particle Data Group Collaboration, R. L. Workman and Others, Review of Particle
651
+ Physics, PTEP 2022 (2022) 083C01.
652
+ – 12 –
653
+
654
+ [53] U. Langenfeld, S. Moch, and P. Uwer, Measuring the running top-quark mass, Phys. Rev. D
655
+ 80 (2009) 054009, [arXiv:0906.5273].
656
+ [54] CMS Collaboration, A. M. Sirunyan et al., Measurement of t¯t normalised multi-differential
657
+ cross sections in pp collisions at √s = 13 TeV, and simultaneous determination of the strong
658
+ coupling strength, top quark pole mass, and parton distribution functions, Eur. Phys. J. C 80
659
+ (2020), no. 7 658, [arXiv:1904.05237].
660
+ [55] O. Akarsu, S. Kumar, E. Özülker, J. A. Vazquez, and A. Yadav, Relaxing cosmological
661
+ tensions with a sign switching cosmological constant: Improved results with Planck, BAO and
662
+ Pantheon data, arXiv:2211.05742.
663
+ [56] O. Akarsu, S. Kumar, E. Özülker, and J. A. Vazquez, Relaxing cosmological tensions with a
664
+ sign switching cosmological constant, Phys. Rev. D 104 (Dec, 2021) 123512.
665
+ [57] O. Akarsu, J. D. Barrow, L. A. Escamilla, and J. A. Vazquez, Graduated dark energy:
666
+ Observational hints of a spontaneous sign switch in the cosmological constant, Phys. Rev. D
667
+ 101 (Mar, 2020) 063528.
668
+ [58] A. Reeves, L. Herold, S. Vagnozzi, B. D. Sherwin, and E. G. M. Ferreira, Restoring
669
+ cosmological concordance with early dark energy and massive neutrinos?, arXiv:2207.01501.
670
+ [59] K. Naidoo, M. Jaber, W. A. Hellwing, and M. Bilicki, A dark matter solution to the H0 and
671
+ σ8 tensions, and the integrated Sachs-Wolfe void anomaly, arXiv:2209.08102.
672
+ [60] R. Abdul Khalek, S. Bailey, J. Gao, L. Harland-Lang, and J. Rojo, Towards Ultimate Parton
673
+ Distributions at the High-Luminosity LHC, Eur. Phys. J. C 78 (2018), no. 11 962,
674
+ [arXiv:1810.03639].
675
+ [61] P. Azzi, L. Gouskos, M. Selvaggi, and F. Simon, Higgs and top physics reconstruction
676
+ challenges and opportunities at FCC-ee, Eur. Phys. J. Plus 137 (2022), no. 1 39,
677
+ [arXiv:2107.05003].
678
+ – 13 –
679
+
ANFKT4oBgHgl3EQfVi5k/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
AtAzT4oBgHgl3EQfF_uW/content/tmp_files/2301.01021v1.pdf.txt ADDED
@@ -0,0 +1,1024 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ A fixed point can hide another one: the nonperturbative behavior of
2
+ the tetracritical fixed point of the O(N) models at large N
3
+ Shunsuke Yabunaka1, ∗ and Bertrand Delamotte2
4
+ 1Advanced Science Research Center, Japan Atomic Energy Agency, Tokai, 319-1195, Japan
5
+ 2Sorbonne Universit´e, CNRS, Laboratoire de Physique Th´eorique de la Mati`ere Condens´ee, LPTMC, F-75005 Paris, France.
6
+ (Dated: January 4, 2023)
7
+ We show that at N = ∞ and below its upper critical dimension, d < dup, the critical and
8
+ tetracritical behaviors of the O(N) models are associated with the same renormalization group fixed
9
+ point (FP) potential. Only their derivatives make them different with the subtleties that taking
10
+ their N → ∞ limit and deriving them do not commute and that two relevant eigenperturbations
11
+ show singularities. This invalidates both the ϵ− and the 1/N− expansions. We also show how the
12
+ Bardeen-Moshe-Bander line of tetracritical FPs at N = ∞ and d = dup can be understood from a
13
+ finite-N analysis.
14
+ Field theories sometimes exhibit nonperturbative fea-
15
+ tures such as confinement [1], presence of bound states [2]
16
+ or exotic excitations [3], fixed points (FPs) of the renor-
17
+ malization group (RG) flows that are nonperturbative as
18
+ in the Kardar-Parisi-Zhang equation [4], divergence of
19
+ the perturbative RG flow at a finite RG scale [5], pres-
20
+ ence of a cusp in the FP potential as in the random field
21
+ Ising model [6], to cite but a few. Very often, these non-
22
+ perturbative effects are assumed either to occur in rather
23
+ complicated theories such as gauge and string theories or
24
+ in highly nontrivial statistical models.
25
+ O(N) models, which are the simplest scalar field theo-
26
+ ries, are often implicitly considered to be immune to these
27
+ complex phenomena.
28
+ Perturbative methods are there-
29
+ fore assumed to work almost all the time for these mod-
30
+ els, the exception to the rule being the Bardeen-Moshe-
31
+ Bander (BMB) phenomenon [7], related to the existence
32
+ of a line of tricritical FPs at N = ∞ and d = 3, which re-
33
+ quires nonperturbative FPs to be fully understood from a
34
+ large-N analysis [8]. From this viewpoint, the enormous
35
+ success of the ϵ = 4 − d expansion for the perturbative
36
+ calculation of the critical exponents associated with the
37
+ Wilson-Fisher (WF) FP [11] could let us believe that the
38
+ critical physics of the O(N) models is fully understood
39
+ for any N and d, especially since it is corroborated by
40
+ the 1/N and ϵ = d − 2 expansions [11].
41
+ Our goal in this Letter is to show instead that although
42
+ the critical physics of the O(N) models, described by the
43
+ WF FP, is fully under perturbative control at both finite
44
+ and infinite N, the tetracritical physics of these models
45
+ at N = ∞ –and probably of infinitely many multicritical
46
+ behaviors– is not. We show below (i) that at N = ∞, it
47
+ is also associated with the WF FP, which is unexpected,
48
+ and (ii) that it nonetheless shows non-perturbative fea-
49
+ tures that are beyond the reach of the standard imple-
50
+ mentation of both the large-N and ϵ- expansions. We
51
+ show in particular a very intriguing phenomenon related
52
+ to the large-N limit of the tetracritical FP of the O(N)
53
+ models: from the second order, the derivatives of the
54
55
+ N = ∞ tetracritical FP potential, that is, of the WF FP
56
+ potential, are not identical to the limit of the derivatives
57
+ of the finite-N tetracritical FP potentials when N → ∞.
58
+ This turns out to be crucial for understanding the large-
59
+ N limit of tetracritical phenomena and shows that this
60
+ limit is much less trivial than what is usually said [9–11].
61
+ The perturbative tetracritical FP corresponds to the
62
+ massless (ϕ2)4 theory, the upper critical dimension of
63
+ which is dup = 8/3. It is found in perturbation theory in
64
+ ϵ = 8/3 − d for all N ≥ 1 and it is three times infrared
65
+ unstable [12]. Calling λ/(384N 3) the coupling in front of
66
+ the dimensionless (ϕ2)4 term, the large-N perturbative
67
+ flow equation for λ reads [13]:
68
+ ∂tλ = −3ϵλ + 9λ2
69
+ 4N + O(N −2).
70
+ (1)
71
+ From Eq. (1), we find that at leading order in N, the
72
+ nontrivial FP solution is λ∗ = 4ϵN/3 from which follows
73
+ that perturbation theory does not allow for a control of
74
+ the large-N limit of the tetracritical FP at fixed ϵ. Only
75
+ the double limit N → ∞ and ϵ → 0 such that the product
76
+ ϵN remains finite can possibly be under control. We come
77
+ back on this point in the following.
78
+ Let us recall that in generic dimensions d < 4, the
79
+ only nontrivial FP found in the standard large-N anal-
80
+ ysis of the O(N) models is the WF FP [14]. Thus, no
81
+ tetracritical FP is found at N = ∞ and d < 8/3 which
82
+ is paradoxical considering that it is perturbatively found
83
+ for all N < ∞ and ϵ > 0.
84
+ We show below that the solution to the paradox above
85
+ lies in the field dependence of the tetracritical FP poten-
86
+ tial whereas it cannot be obtained from its field expansion
87
+ and in particular from λ∗. The recourse to functional RG
88
+ methods is therefore mandatory.
89
+ The best way to implement functional RG is to con-
90
+ sider Wilson’s RG, as it is inherently functional [15]. We
91
+ recall below the take-away philosophy of the modern ver-
92
+ sion of Wilson’s RG known as the nonperturbative – or
93
+ functional – renormalization group (NPRG).
94
+ NPRG is based on the idea of integrating fluctuations
95
+ step by step [16]. It is implemented on the Gibbs free
96
+ energy Γ [17–23] of a model defined by an Hamiltonian
97
+ arXiv:2301.01021v1 [cond-mat.stat-mech] 3 Jan 2023
98
+
99
+ 2
100
+ (or euclidean action) H and a partition function Z. To
101
+ this model is associated a one-parameter family of models
102
+ with Hamiltonians Hk = H + ∆Hk and partition func-
103
+ tions Zk, where k is a momentum scale. In Hk, ∆Hk is
104
+ chosen such that only the rapid fluctuations in the origi-
105
+ nal model, those with wavenumbers |q| > k, are summed
106
+ over in the partition function Zk. Thus, the slow modes
107
+ (|q| < k) need to be decoupled in Zk and this is achieved
108
+ by giving them a mass of order k, that is by taking for
109
+ ∆Hk a quadratic (mass-like) term, which is nonvanishing
110
+ only for the slow modes:
111
+ Zk[J] =
112
+
113
+ Dϕi exp(−H[ϕ] − ∆Hk[ϕ] + J · ϕ)
114
+ (2)
115
+ with ∆Hk[ϕ] =
116
+ 1
117
+ 2
118
+
119
+ q Rk(q2)ϕi(q)ϕi(−q), where, for in-
120
+ stance, Rk(q2) = (k2 − q2)θ(k2 − q2) and J · ϕ =
121
+
122
+ x Ji(x)ϕi(x). The k-dependent Gibbs free energy Γk[φ]
123
+ is defined as the (slightly modified) Legendre transform
124
+ of log Zk[J]:
125
+ Γk[φ] + log Zk[J] = J · φ − 1
126
+ 2
127
+
128
+ q
129
+ Rk(q2)φi(q)φi(−q) (3)
130
+ with
131
+
132
+ q =
133
+
134
+ ddq/(2π)d.
135
+ With the choice of regulator
136
+ function Rk above, Γk[φ] interpolates between the Hamil-
137
+ tonian H when k is of order of the ultraviolet cut-off Λ
138
+ of the theory: ΓΛ ∼ H, and the Gibbs free energy Γ of
139
+ the original model when k = 0: Γk=0 = Γ. The exact
140
+ RG flow equation of Γk gives the evolution of Γk with
141
+ k between these two limiting cases. It is known as the
142
+ Wetterich equation. It reads [18]:
143
+ ∂tΓk[φ] = 1
144
+ 2Tr[∂tRk(q2)(Γ(2)
145
+ k [q, −q; φ] + Rk(q))−1], (4)
146
+ where t = log(k/Λ), Tr stands for an integral over q and
147
+ a trace over group indices and Γ(2)
148
+ k [q, −q; φ] is the matrix
149
+ of the Fourier transforms of δ2Γk/δφi(x)δφj(y).
150
+ In most cases, Eq. (4) cannot be solved exactly and
151
+ approximations are mandatory. The best known approx-
152
+ imation consists in expanding Γk in powers of the deriva-
153
+ tives of φi and to truncate the expansion at a given fi-
154
+ nite order[24–32]. The approximation at lowest order is
155
+ dubbed the local potential approximation (LPA). For the
156
+ O(N) model it consists in approximating Γk by:
157
+ Γk[φ] =
158
+
159
+ x
160
+ �1
161
+ 2(∇φi)2 + Uk(φ)
162
+
163
+ (5)
164
+ where φ = √φiφi. Fixed points are found only for di-
165
+ mensionless quantities and the standard large-N limit
166
+ by rescaling the field and the potential by factors N −1/2
167
+ and N −1 respectively.
168
+ Thus, we define the dimen-
169
+ sionless and rescaled field ¯φ and potential ¯Uk as ¯φ =
170
+ v
171
+ − 1
172
+ 2
173
+ d
174
+ k
175
+ 2−d
176
+ 2 N −1/2φ and ¯Uk(¯φ) = v−1
177
+ d k−dN −1Uk (φ) with
178
+ v−1
179
+ d
180
+ = 2d−1dπd/2Γ( d
181
+ 2). The LPA flow of ¯Uk then reads:
182
+ ∂t ¯Uk(¯φ) = − d ¯Uk(¯φ) + 1
183
+ 2(d − 2)¯φ ¯U ′
184
+ k(¯φ)+
185
+
186
+ 1 − 1
187
+ N
188
+
189
+ ¯φ
190
+ ¯φ + ¯U ′
191
+ k(¯φ) + 1
192
+ N
193
+ 1
194
+ 1 + ¯U ′′
195
+ k (¯φ)
196
+ (6)
197
+ FIG. 1. d = 2.6: ¯U(¯φ) for the T3 FP of Eq. (6). Green, red,
198
+ blue and black curves correspond to N = 1500, 2250, 4500
199
+ and 42000. The orange dashed curve corresponds to the WF
200
+ FP at N = ∞. Inset: Close view of ¯U(¯φ) around ¯φi.
201
+ with ∂t = k∂k. The standard large-N limit of the LPA
202
+ flow equation above is obtained by (i) replacing the fac-
203
+ tor 1 − 1/N by 1, (ii) dropping the last term in Eq. (6)
204
+ because it is assumed to be sub-leading [33]. As a con-
205
+ sequence of the two steps above, the explicit dependence
206
+ in N in Eq. (6) disappears in the large-N limit.
207
+ The crucial point of the large-N limit is that assuming
208
+ point (ii) above, the resulting LPA flow equation on ¯Uk
209
+ can be shown to be exact in the limit N → ∞ [34]. Under
210
+ this assumption, all FPs of the O(N) models have been
211
+ found exactly at N = ∞ [14, 33–36]. The result is the
212
+ following: In a generic dimension d < 4 there is only one
213
+ nongaussian FP at N = ∞ which is the usual Wilson-
214
+ Fisher FP (WF). The exceptions to the rule above are the
215
+ BMB lines of FPs [7, 14, 37–39] existing in dimensions
216
+ d = 2 + 2/p with p an integer larger than 1.
217
+ We now show that the procedure described above is too
218
+ restrictive to study the large-N limit of the tetracritical
219
+ FPs. As said above, the standard large-N analysis con-
220
+ sists in neglecting the last term in Eq. (6). However, this
221
+ term is negligible only if (1 + ¯U ′′
222
+ k (¯φ))−1 does not coun-
223
+ terbalance at large N its 1/N prefactor for some finite
224
+ values of ¯φ. We now show that because of singularities in
225
+ the third derivative of ¯Uk(¯φ), the contribution of the last
226
+ term in Eq. (6) cannot be neglected in the FP equation
227
+ of ¯U ′′
228
+ k (¯φ) obtained by differentiating twice Eq. (6) (see
229
+ footnote below Eq. (8) for more detail). This turns out
230
+ to be sufficient to invalidate the standard large-N limit
231
+ in the tetracritical case.
232
+ We have numerically solved Eq. (6) and have found
233
+ for several values of N and d < 8/3 the perturbative
234
+ tetracritical FP that we call T3(N, d). As expected, T3
235
+ bifurcates from the Gaussian FP in d = 8/3−. We have
236
+ followed it down to d = 2.6, see Fig.
237
+ 1 and Fig.
238
+ 3
239
+ of the Suppl. Mat. The FP potential of T3, (i) shows
240
+ as expected two maxima, one of which being located at
241
+ ¯φ = 0 and another one at ¯φ2 > 0, and two minima at
242
+ ¯φ1 and ¯φ3 such that ¯φ3 > ¯φ2 > ¯φ1 > 0, see Fig.
243
+ 1,
244
+ (ii) can be continuously followed up to arbitrarily large
245
+ values of N at fixed d < 8/3, (iii) has its three extrema
246
+
247
+ T()
248
+ 0.38466
249
+ 0.38464
250
+ 0.9
251
+ 0.38462
252
+ 0.8
253
+ 0.38460
254
+ 0.7
255
+ 1.75
256
+ 1.80
257
+ 1.85
258
+ $1.90
259
+ 0.6
260
+ 0.5
261
+ 2.03
262
+ ¯φ1, ¯φ2, ¯φ3 approaching each other when N is increased at
263
+ fixed d. These extrema tend to a common value ¯φ0 when
264
+ N → ∞ which is the minimum of the FP potential, see
265
+ Fig. 1 and Fig. 4 of the Suppl. Mat. Point (ii) above is
266
+ paradoxical because it seems to contradict the standard
267
+ large-N approach where only the WF FP is found in a
268
+ generic dimension d < 8/3 at N = ∞. We now show
269
+ that the WF FP potential at N = ∞ is in fact the limit
270
+ when N → ∞ of the potential of T3 for d < 8/3. This
271
+ solves the above paradox because it explains why on one
272
+ hand there exists a nontrivial tetracritical FP at N = ∞
273
+ and d < 8/3 and on the other hand that there is no
274
+ other nontrivial and smooth solution of Eq. (6) at N =
275
+ ∞ than the WF FP potential. However, this creates a
276
+ new paradox since obviously the critical and tetracritical
277
+ universal behaviors cannot be the same since the two FPs
278
+ do not have the same number of unstable eigendirections.
279
+ We now explain in detail this new paradox.
280
+ We can see on Fig. 1 that the FP potentials found
281
+ in d = 2.6 for large values of N are extremely flat in
282
+ the region, ¯φ ∈ [¯φ1, ¯φ3] because the three extrema are
283
+ very close and the height of the barrier between the two
284
+ minima very small. We have numerically found that the
285
+ height of the barrier scales as N −1 and the distance be-
286
+ tween the two minima as N −1/2 so that the curvatures
287
+ ¯U ′′(¯φi) at the three extrema approach constant values as
288
+ N → ∞, see Fig. 4 of the Suppl. Mat. This suggests
289
+ that ¯U ′′(¯φ) while being well-behaved everywhere but be-
290
+ tween the three extrema, changes very rapidly within a
291
+ boundary layer around ¯φ0 of typical width N −1/2, mak-
292
+ ing divergent ¯U ′′′(¯φ0) when N → ∞.
293
+ It is not common in physics to encounter this kind
294
+ of situation where a series of functions fn(x) tends to a
295
+ smooth function f∞(x) whereas from a certain order p,
296
+ their derivatives f (p)
297
+ n (x) do not tend to f (p)
298
+ ∞ (x). However,
299
+ a simple toy model explains trivially how this can occur.
300
+ Consider the series of functions fn(x) = n−1 sin(n2x).
301
+ Obviously, f∞(x) ≡ 0 which implies that f ′
302
+ ∞(x) ≡ 0
303
+ whereas limn→∞ f ′
304
+ n(0) = ∞.
305
+ In our case, at fixed d < 8/3, the limit of the T3 po-
306
+ tentials when N → ∞ is a nontrivial and well-defined
307
+ function that therefore must be the WF FP potential.
308
+ We have checked that it is indeed the limit of T3 when
309
+ N → ∞, see Fig. 1. The difference between the critical
310
+ and tetracritical behaviors is therefore not visible on the
311
+ potentials themselves but only on their derivatives as we
312
+ now show.
313
+ Let us study the boundary layer around ¯φ0. It is con-
314
+ venient for what follows to change variables. Following
315
+ Ref. [40], we define: V (µ) = U(φ) + (φ − Φ)2/2 with
316
+ µ = Φ2 and φ − Φ = −2ΦV ′(µ). As above, it is conve-
317
+ nient to rescale µ and V (µ): ¯µ = µ/N, ¯V = V/N. In
318
+ terms of these quantities, the FP equation for ¯V (¯µ) reads
319
+ 0 = 1 − d ¯V + (d − 2)¯µ ¯V ′ + 4¯µ ¯V ′2 − 2 ¯V ′ − 4
320
+ N ¯µ ¯V ′′. (7)
321
+ Eq. (7) has two remarkable features: (i) it is much sim-
322
+ pler than Eq. (6) because the nonlinearity comes only
323
+ 0
324
+ 2
325
+ 4
326
+ 6
327
+ 8
328
+ 10
329
+ 12
330
+ 14 μ
331
+ 0.02
332
+ 0.04
333
+ 0.06
334
+ 0.08
335
+ V''[μ]
336
+ N=6×103
337
+ N=1.7×104
338
+ N=3.2×106
339
+ N=∞WF
340
+ FIG. 2. Second derivative of the WF and T3 FP potentials
341
+ for different values of N in d = 2.6.
342
+ from the ( ¯V ′)2 term, (ii) it is the LPA equation obtained
343
+ from the Wilson-Polchinski (WP) version of the NPRG
344
+ [15, 41, 42]. Thus, ¯V (¯µ) is related to the potential ¯U(¯φ) of
345
+ the Wetterich version of the RG by the Legendre trans-
346
+ form of Eq.
347
+ (3).
348
+ The standard large-N analysis per-
349
+ formed in this version of the NPRG consists here again
350
+ in neglecting the last term in Eq. (7) because it is sup-
351
+ pressed by a 1/N factor. Under the assumption that this
352
+ term is indeed negligible, the resulting equation can be
353
+ solved exactly in the large-N limit [14, 35]. However, at
354
+ large N, it is clear on Eq. (7) that we have to deal with
355
+ singular perturbation theory since the small parameter
356
+ used in the 1/N expansion is in front of the term of high-
357
+ est derivative, that is, ¯V ′′. In this case, it is well-known
358
+ that at large N a boundary layer can exist for a partic-
359
+ ular value of ¯µ that becomes a singularity at N = ∞,
360
+ making this term non negligible [43].
361
+ The value of ¯µ corresponding to ¯φ0 is called ¯µ0 and
362
+ is the minimum of ¯V (¯µ) at N
363
+ = ∞.
364
+ We find for
365
+ ¯V (¯µ) the same features about its three extrema ¯µi as
366
+ for ¯U(¯φ) at ¯φi: The three extrema ¯µi approach each
367
+ other and to ¯µ0 as N → ∞, the distances between
368
+ them scale as N −1/2 and the curvatures ¯V ′′(¯µi) as N 0.
369
+ Taking into account the scaling around ¯µ0 inside the
370
+ boundary layer, we introduce another scaled variable
371
+ ˜µ = N 1/2(¯µ − ¯µ0).
372
+ Since at N = ∞, ¯V ′(¯µ) vanishes
373
+ at ¯µ = ¯µ0, ¯V (¯µ0) should approach 1/d at leading order
374
+ in N −1/2. We therefore define a scaled boundary layer
375
+ by ˜VN(˜µ) = N
376
+ � ¯V (¯µ0 + N −1/2˜µ) − 1/d
377
+
378
+ which implies
379
+ ˜V ′′
380
+ N(˜µ) = ¯V ′′(¯µ0 + N −1/2˜µ). We plot ˜V ′′
381
+ N(˜µ) for several
382
+ values of N in Fig. 5 of the Suppl. Mat.
383
+ By substituting ˜VN(˜µ) by its value in Eq.
384
+ (7) and
385
+ solving it at order O(N −1/2), we find that ¯µ0 = 2/(d−2).
386
+ At order O(N −1), Eq. (7) becomes
387
+ − 8 ˜V ′′
388
+ ∞(˜µ)
389
+ d − 2 + 8 ˜V ′
390
+ ∞(˜µ)2
391
+ d − 2
392
+ +(d−2)˜µ ˜V ′
393
+ ∞(˜µ)−d ˜V∞(˜µ) = 0 (8)
394
+ [44] which is clearly invariant under ˜µ → −˜µ from which
395
+ it follows that ˜V ′
396
+ ∞(0) = 0. At ˜µ = ∞, ˜V ′′
397
+ ∞(˜µ) should tend
398
+ to a finite value that matches with ¯V ′′(µ) at ¯µ+
399
+ 0 . This
400
+ implies that the solution of Eq. (8) should be quadratic
401
+ when ˜µ → ∞. Substituting ˜V∞(˜µ) by ˜V ′′
402
+ ∞(˜µ = ∞)˜µ2/2
403
+ in Eq. (8) and balancing the leading terms as ˜µ → ∞,
404
+ we find that ˜V ′′
405
+ ∞(˜µ = ∞) = (−d2 + 6d − 8)/16. Imposing
406
+
407
+ 4
408
+ the two boundary conditions found above at ˜µ = 0 and
409
+ ˜µ = ∞ selects a unique and globally defined solution
410
+ ˜V ′′
411
+ ∞(˜µ) of Eq. (8) shown in Fig. 5 of the Suppl. Mat.
412
+ We find ¯V ′′(¯µ+
413
+ 0 ) = ¯V ′′
414
+ WF(¯µ0) = ˜V ′′
415
+ ∞(˜µ = ∞) which proves
416
+ the matching at N = ∞ between the boundary layer and
417
+ the potential outside of the layer, see Fig. 2. We have
418
+ shown in Fig. 6 of the Suppl. Mat. the boundary layer
419
+ for ¯U ′′(¯φ) analogous to that of ¯V ′′(¯µ). To conclude, we
420
+ have proven that for d < 8/3, a boundary layer develops
421
+ at large N for the second derivative of the T3 potential
422
+ that becomes a singularity when N → ∞. What remains
423
+ to be understood is its physical relevance.
424
+ At first sight, what we have obtained for T3 looks para-
425
+ doxical because we could think that its potential being
426
+ identical to the WF potential at N = ∞, the linearized
427
+ flow around these two FPs should also be identical and
428
+ thus the same for all critical exponents. We now show
429
+ that this naive argument is wrong.
430
+ We have computed in d < 8/3 the relevant eigenvalues
431
+ of the RG flow around T3 and WF at finite and large N
432
+ and as expected we have found three for T3 and one for
433
+ WF. When N → ∞, one of the three eigenvalues at T3
434
+ tends as expected to d − 2 which is the relevant eigen-
435
+ value ν−1 of the critical WF FP at N = ∞ [11, 14]. The
436
+ nontrivial point is that the two other relevant eigenvalues
437
+ at T3 have a well-defined limit when N → ∞ although
438
+ they do not play any role for the critical behavior of the
439
+ O(N = ∞) model. The solution to this paradox is that
440
+ they are associated with eigenperturbations that become
441
+ singular when N → ∞. That these two eigenperturba-
442
+ tions become singular is clear for one of them, called δ ¯V2,
443
+ on Fig. 9 of the Suppl. Mat. As for the other one, δ ¯V1,
444
+ its slope at ¯µ0 diverges as N 1/3 which implies that at
445
+ N = ∞, it becomes discontinuous at ¯µ0, see Figs. 9 and
446
+ 10 of the Suppl. Mat. For ordinary second order phase
447
+ transitions, these eigenperturbations are excluded which
448
+ explains that the associated relevant eigenvalues do not
449
+ play any role. This solves all the paradoxes associated
450
+ with the tetracritical FPs at N = ∞ and d < 8/3.
451
+ What remains to be studied is the particular case N =
452
+ ∞ and d = 8/3 where a line, called the BMB line, of
453
+ smooth tetracritical FPs shows up. It is obtained in the
454
+ WP version of the RG by integrating Eq. (7) in which
455
+ the last term, proportional to 1/N, has been discarded.
456
+ It is given by the following implicit expression [14]:
457
+ ¯µ± =
458
+ C
459
+ ¯V ′ �
460
+ 1 − 2 ¯V ′�
461
+ � ±2 ¯V ′
462
+ 1 − 2 ¯V ′
463
+ �4/3
464
+ + 2f(4 ¯V ′),
465
+ (9)
466
+ where f(x), which is analytic for x < 2, is given by
467
+ f(x) =
468
+ 3
469
+ 2 − x +
470
+ 4x
471
+ (2 − x)7/3
472
+ � 1
473
+ 0
474
+ dz
475
+ �2 − xz
476
+ z
477
+ �1/3
478
+ (10)
479
+ and ¯µ± correspond to the two branches ¯µ > 3 and ¯µ < 3,
480
+ respectively. The derivative of the potential ¯V ′ is positive
481
+ (negative) on the former (latter) branch and C is a non-
482
+ negative integration constant.
483
+ ¯V (¯µ) is analytic at ¯µ =
484
+ ¯µ0 = 3 and ¯V ′(¯µ = 3) = 0.
485
+ In Fig.
486
+ 7 of the Suppl.
487
+ Mat. different ¯V ′(¯µ) corresponding to different FPs of
488
+ the BMB line are shown. All FPs along the BMB line
489
+ share the same critical exponents, that is, the exponents
490
+ of the Gaussian FP which is itself tetracritical. Notice
491
+ that the WF FP which corresponds to C = 0, is the end
492
+ point of this line and deserves special attention. We come
493
+ back on this point in the following.
494
+ From Eq. (1), we have seen that λ∗ remains constant at
495
+ leading order in 1/N along the hyperbola of constant ϵN
496
+ of the (d, N) plane. This suggests that when the double
497
+ limit d → 8/3 and N → ∞ is taken at fixed α = ϵN, T3
498
+ converges in d = 8/3 to one of the FPs of the BMB line.
499
+ We have analytically and numerically checked this and
500
+ have derived analytically the relation between α and C:
501
+ α = 162/C3, see Suppl. Mat. and Fig. 11.
502
+ Two extreme cases are worth studying.
503
+ First, the
504
+ Gaussian FP corresponds to the limit N → ∞ at fixed
505
+ dimension d = 8/3, that is, at α = 0. It corresponds
506
+ to C = ∞ in Eq. (9). Second, α = ∞, which implies
507
+ C = 0, corresponds to taking the limit ϵ → 0 at fixed
508
+ N = ∞, that is, to following the WF FP at N = ∞ up
509
+ to d = 8/3. However, at finite ϵ and N = ∞, we know
510
+ from the analysis above that the last term in Eq. (7) can-
511
+ not be neglected. Consistently, the same occurs for the
512
+ BMB line: the WF FP potential is indeed the end point
513
+ of the BMB line obtained by taking the limit C → 0 in
514
+ Eq. (9) but the derivatives of this potential can only be
515
+ studied by retaining the last term in Eq. (7). Here again,
516
+ this explains why the T3 FP in the C → 0 limit is three
517
+ times unstable and not only once unstable.
518
+ To conclude, we have solved the paradox of the appar-
519
+ ent absence of a nontrivial tetracritical FP at N = ∞
520
+ and d < 8/3 by showing that this FP does exist but is
521
+ nothing else than the WF FP up to the subtlety that
522
+ the derivatives of the tetracritical FP potential are not
523
+ the derivatives of the WF FP potential. This makes the
524
+ large-N limit of the O(N) model much less trivial than
525
+ is usually advocated at least for multicritical phenom-
526
+ ena.
527
+ The fact that the tetracritical FP has two more
528
+ unstable infrared directions than the WF FP is related
529
+ to this subtle point because they are associated with sin-
530
+ gular eigenperturbations, a possibility which is usually
531
+ not considered. We conjecture that what has been found
532
+ above at large N and for d ≤ 8/3 is valid for all mul-
533
+ ticritical points with an odd number of eigendirections
534
+ below or at their upper critical dimension because the
535
+ BMB lines for all of them terminate at the WF FP [14],
536
+ a fact that in itself is almost enough to imply everything
537
+ else. Let us finally point out that what we have found for
538
+ the tetracritical FP is very different from what was found
539
+ around d = 3 at large-N in the tricritical case which re-
540
+ quired the existence of new FPs to be fully understood
541
+ at finite N [45–48]. We also conjecture that this phe-
542
+ nomenon is not specific to the O(N) models but should
543
+ rather be generic.
544
+ We acknowledge A. Codello and N. Defenu and for
545
+ correspondence and discussions at an early stage of this
546
+
547
+ 5
548
+ work. S. Y. was supported by Grant-in-Aid for Young
549
+ Scientists (18K13516).
550
+ [1] M. Pel´aez, U. Reinosa, J. Serreau, M. Tissier, and N.
551
+ Wschebor, Rept. Prog. Phys. 84, 124202 (2021).
552
+ [2] P. Hoyer, Journey to the Bound States , Springer (2021).
553
+ [3] X.-Y. Guo, Y. Heo, and M.F.M. Lutz, Phys. Lett. B 791,
554
+ 86 (2019)
555
+ [4] L. Canet, H. Chat´e, B. Delamotte, and N. Wschebor,
556
+ Phys. Rev. E 84, 061128 (2011); Phys. Rev. E 86,
557
+ 019904(E) (2012); T. Kloss, L. Canet, B. Delamotte, and
558
+ N. Wschebor, Phys. Rev. E 89, 022108 (2014).
559
+ [5] H.-K. Janssen, F. van Wijland, O. Deloubri`ere, and U.C.
560
+ T¨auber, Phys. Rev. E 70, 056114 (2004); D. Gredat, H.
561
+ Chat´e, B. Delamotte, and I. Dornic, Phys. Rev. E 89,
562
+ 010102(R) (2014).
563
+ [6] M. Tissier and G. Tarjus, Phys. Rev. B 78, 024204
564
+ (2008); Phys. Rev. Lett. 107, 041601 (2011); Phys. Rev.
565
+ B 85, 104202 (2012); Phys. Rev. B 85, 104203 (2012).
566
+ [7] W. A. Bardeen, M. Moshe, and M. Bander, Phys. Rev.
567
+ Lett. 52, 1188 (1984).
568
+ [8] C. Fleming, B. Delamotte, and S. Yabunaka, Phys. Rev.
569
+ D 102, 065008 (2020).
570
+ [9] T. H. Berlin, and M. Kac, Phys. Rev. 86, 821 (1952); H.
571
+ E. Stanley, Phys. Rev. 176, 718 (1968); E. Br´ezin and
572
+ D. J. Wallace, Phys. Rev. B 7, 1967 (1973); S.k. Ma, J.
573
+ Math. Phys. 15, 1866 (1974).
574
+ [10] G. Eyal, M. Moshe, S. Nishigaki, and J. Zinn-Justin,
575
+ Nucl. Phys. B 470 [FS], 369 (1996).
576
+ [11] J. Zinn-Justin, Quantum field theory and critical phe-
577
+ nomena Fourth Edition, Oxford University Press, (2002).
578
+ [12] C. Itzykson and J. Drouffe, Statistical Field Theory:
579
+ Volume 1, From Brownian Motion to Renormalization
580
+ and Lattice Gauge Theory, Cambridge University Press,
581
+ (1989).
582
+ [13] N. Defenu and A. Codello, arXiv:2005.10827.
583
+ [14] J. Comellas and A. Travesset, Nucl. Phys. B 498, 539
584
+ (1997).
585
+ [15] K. G. Wilson and J. Kogut, Phys. Rep., Phys. Lett.,12C
586
+ (1974) 75.
587
+ [16] K. G. Wilson, Phys. Rev. B 4, 3174 (1971).
588
+ [17] C. Wetterich, Nucl. Phys. B 352, 529 (1991).
589
+ [18] C. Wetterich, Phys. Lett. B 301, 90 (1993).
590
+ [19] U. Ellwanger, Z. Phys. C. Part. Fields 58, 619 (1993).
591
+ [20] T. R. Morris, Int. J. Mod. Phys. A 09, 2411 (1994).
592
+ [21] J. Berges, N. Tetradis and C. Wetterich, Phys. Rept. 363,
593
+ 223 (2002)
594
+ [22] B. Delamotte, D. Mouhanna, and M. Tissier, Phys. Rev.
595
+ B 69, 134413 (2004).
596
+ [23] B. Delamotte, “An Introduction to the nonperturba-
597
+ tive renormalization group”, Lect. Notes Phys. 852, 49
598
+ (2012).
599
+ [24] L. Canet, B. Delamotte, D. Mouhanna, and J. Vidal,
600
+ Phys. Rev. B 68, 064421 (2003).
601
+ [25] L. Canet, Phys. Rev. B 71, 012418 (2005).
602
+ [26] T. Kloss, L. Canet, B. Delamotte, and N. Wschebor,
603
+ Phys. Rev. E 89, 022108 (2014); L. Canet, H. Chat´e, B.
604
+ Delamotte, and N. Wschebor, Phys. Rev. E 84, 061128
605
+ (2011); Phys. Rev. E 86, 019904(E) (2012).
606
+ [27] B. Delamotte and L. Canet, Condensed Matter Phys. 8,
607
+ 163 (2005).
608
+ [28] F. Benitez, R. M´endez-Galain, and N. Wschebor, Phys.
609
+ Rev. B 77, 024431 (2008).
610
+ [29] L. Canet, B. Delamotte, O. Deloubri`ere, and N. Wsche-
611
+ bor, Phys. Rev. Lett. 92, 195703 (2004); L. Canet, H.
612
+ Chat´e, and B. Delamotte, Phys. Rev. Lett. 92, 255703
613
+ (2004); L. Canet, B. Delamotte, D. Mouhanna, and J.
614
+ Vidal, Phys. Rev. D 67, 065004 (2003); L. Canet, H.
615
+ Chat´e, B. Delamotte, I. Dornic, and M. A. Mu˜noz, Phys.
616
+ Rev. Lett. 95, 100601 (2005).
617
+ [30] L. Canet, B. Delamotte, and N. Wschebor, Phys. Rev. E
618
+ 93, 063101 (2016).
619
+ [31] F. L´eonard and B. Delamotte, Phys. Rev. Lett. 115,
620
+ 200601 (2015).
621
+ [32] Balog, Ivan and Chat´e, Hugues and Delamotte, Bertrand
622
+ and Marohni´c, Maroje and Wschebor, Nicol´as, Phys.
623
+ Rev. Lett. 123, 240604 (2019)
624
+ [33] N. Tetradis and D. F. Litim, Nucl. Phys. B 464, 492
625
+ (1996).
626
+ [34] M. D’Attanasio and T. R. Morris, Phys. Lett. B 409, 363
627
+ (1997).
628
+ [35] Y. Kubyshin, R. Neves, and R. Potting, Int. J. Mod.
629
+ Phys. A 16, 2065 (2001).
630
+ [36] A. Katsis and N. Tetradis, Phys. Lett. B 780, 491-494
631
+ (2018).
632
+ [37] F. David, D. A. Kessler, and H. Neuberger, Phys. Rev.
633
+ Lett. 53, 2071 (1984).
634
+ [38] H. Omid, G. W. Semenoff, and L. C.R. Wijewardhana,
635
+ Phys. Rev. D 94, 125017 (2016).
636
+ [39] D. F. Litim, E. Marchais, and P. Mati, Phys. Rev. D 95,
637
+ 125006 (2017).
638
+ [40] T. R. Morris, J. High Energy Phys., 07, 027 (2005).
639
+ [41] J. Polchinski, Nucl. Phys. B 231 (1984) 269.
640
+ [42] A. Hasenfratz and P. Hasenfratz, Nucl. Phys. B, 270, 687
641
+ (1986).
642
+ [43] M. H. Holmes, Introduction to perturbation methods Sec-
643
+ ond edition, Springer (2012).
644
+ [44] Note that the first term in Eq. (8) comes from the last
645
+ term in Eq. (6) or Eq. (7), which is formally proportional
646
+ to N −1 and neglected in the usual large-N analysis. How-
647
+ ever this term is indispensable to describe the boundary
648
+ layer of ¯U ′′(¯φ) or ¯V ′′(¯µ).
649
+ [45] R. D. Pisarski, Phys. Rev. Lett. 48, 574 (1982).
650
+ [46] H. Osborn and A. Stergiou, J. High Energy Phys. 5 51
651
+ (2018).
652
+ [47] S. Yabunaka and B. Delamotte, Phys. Rev. Lett. 119,
653
+ 191602 (2017); Phys. Rev. Lett. 121, 231601 (2018).
654
+ [48] S. Yabunaka, C. Fleming, and B. Delamotte, Phys. Rev.
655
+ E 106, 054105 (2022).
656
+
657
+ 6
658
+ SUPPLEMENTAL MATERIALS
659
+ I.
660
+ T3 FP POTENTIALS IN d < 8/3
661
+ We show in Fig. 3 the tetracritical FP potential ¯U(¯φ)
662
+ obtained with the LPA and solution of Eq. (6) for small
663
+ values of N. They have the typical shape of a tetracritical
664
+ potential showing two nontrivial minima.
665
+ 0
666
+ 2
667
+ 4
668
+ 6
669
+ 8 ϕ
670
+ 0.375
671
+ 0.38
672
+ 0.385
673
+ 0.39
674
+ U(ϕ)
675
+ N=4.5
676
+ N=1
677
+ FIG. 3. ¯U(¯φ) for the T3 FP for different values of N in d = 2.6.
678
+ II.
679
+ LARGE-N BEHAVIOR OF THE EXTREMA
680
+ OF THE TETRACRITICAL POTENTIAL
681
+ The three nontrivial extrema of the T3 FP potential
682
+ in either the WP or Wetterich version of the RG, shown
683
+ in Fig. 1 of the main text, behave the same way when
684
+ N → ∞. We show on Fig. 4 the scaling in N of the height
685
+ of the barrier between the extrema ¯φi of the rescaled
686
+ potential ¯U(¯φ) of the Wetterich version of the RG, as
687
+ well as the distance between them. These extrema are
688
+ shown in Fig. 1 of the main text.
689
+ 2500
690
+ 3500
691
+ 4500N
692
+ 6.0×10-6
693
+ 8.0×10-6
694
+ 1.0×10-5
695
+ 1.2×10-5
696
+ U[ϕ2]-U[ϕ3]
697
+ 2500
698
+ 3500
699
+ 4500N
700
+ 0.0325
701
+ 0.0350
702
+ 0.0375
703
+ 0.0400
704
+ 0.0425
705
+ 0.0450
706
+ 0.0475
707
+ ϕ3-ϕ2
708
+ FIG. 4. Left: Height of the potential barrier for the T3 FP
709
+ of Eq. (6) for large values of N in d = 2.6 (blue dots). The
710
+ equation of the full line is y = 0.0257/N. Right: Distance
711
+ between the maximum ¯φ2 and the minimum ¯φ3 for the T3 FP
712
+ of Eq. (6) for large values of N in d = 2.6 (blue dots). The
713
+ equation of the full line is y = 2.12506/N 1/2.
714
+ Since the height of the barrier, ∆ ¯U, scales as N −1
715
+ and the distance between the extrema, ∆¯φ, as N −1/2, a
716
+ simple dimensional argument shows that the curvatures
717
+ at these extrema that goes as ∆ ¯U/(∆¯φ)2, do not scale
718
+ with N, that is, tend to constants when N → ∞, a fact
719
+ that we have numerically checked.
720
+ Thus, for d < 8/3
721
+ and at large and finite N, the curvature of ¯U(¯φ) varies
722
+ between a positive value at ¯φ1, a negative value at ¯φ2
723
+ and again a positive value at ¯φ3 on a distance of order
724
+ N −1/2.
725
+ III.
726
+ THE SCALED BOUNDARY LAYER ˜V ′′(˜µ)
727
+ By translating and rescaling by a factor N 1/2 the po-
728
+ sition and the width of the boundary layer of the second
729
+ derivative of the potential ¯V , it is possible to obtain a
730
+ finite limit for this scaled boundary layer when N → ∞.
731
+ We thus define the scaled variable ˜µ = N 1/2(¯µ − ¯µ0)
732
+ where ¯µ0 is the location of the boundary layer and the
733
+ scaled potential by ˜VN(˜µ) = N
734
+ � ¯V (N −1/2˜µ + ¯µ0) − 1/d
735
+
736
+ .
737
+ It follows from the definitions above that ˜V ′′
738
+ N(˜µ) =
739
+ ¯V ′′(¯µ0+N 1/2˜µ). We show in Fig. 5 this scaled boundary
740
+ layer for different values of N at large N as well as its
741
+ limit ˜V ′′
742
+ ∞(˜µ) at N = ∞.
743
+ -1000
744
+ -500
745
+ 500
746
+ 1000
747
+ μ˜
748
+ 0.02
749
+ 0.04
750
+ 0.06
751
+ 0.08
752
+ 0.10
753
+ 0.12
754
+ 0.14
755
+ N=2.5×104
756
+ N=1.7×105
757
+ N=3.2×106
758
+ ˜V''[μ]
759
+ ˜
760
+ FIG. 5. The scaled boundary layer for the second derivative
761
+ of the T3 FP potential ˜V ′′
762
+ N(˜µ), Eqs. (7) for large values of N in
763
+ d = 2.6. The dashed curve is the global solution ˜V ′′
764
+ ∞(˜µ) of Eq.
765
+ (8) at N = ∞. The red horizontal line is y = (−d2+6d−8)/16
766
+ for d = 2.6. It coincides with ˜V ′′
767
+ ∞(˜µ = ∞).
768
+ Notice that a finite difference ¯µ − ¯µ0 translates into an
769
+ infinite ˜µ when N → ∞. The matching at N = ∞ be-
770
+ tween the scaled boundary layer and the value of ¯V ′′(¯µ)
771
+ outside of the layer therefore requires that ˜V ′′
772
+ ∞(∞) =
773
+ ¯V ′′(¯µ+
774
+ 0 ) = (−d2 + 6d − 8)/16 which is the case with the
775
+ solution for the scaled boundary layer given in the main
776
+ text, see also Fig. 5.
777
+ IV.
778
+ THE BOUNDARY LAYER OF ¯U ′′(¯φ)
779
+ The boundary layer has been derived in the main text
780
+ in WP version of the RG because it is simpler in this
781
+ version than in Wetterich version. However, it can also
782
+ be derived directly in this latter version or, once it is
783
+
784
+ 7
785
+ obtained in one version, it can be translated in the other
786
+ by performing the Legendre transform given in Eq. (3).
787
+ 0.5
788
+ 1.0
789
+ 1.5
790
+ 2.0
791
+ 2.5
792
+ ϕ
793
+ 5
794
+ 10
795
+ 15
796
+ U''[ϕ]
797
+ N=1.5⨯104
798
+ N=4.2⨯105
799
+ N=4.2⨯106
800
+ WF N=∞
801
+ FIG. 6. The second derivative of the T3 FP potential ¯U ′′(¯φ) in
802
+ the Wetterich version of the RG, Eq. (6), for different values
803
+ of N in d = 2.6.
804
+ We show in Fig. 6 the boundary layer of ¯U ′′(¯φ) for
805
+ different values of N at large N.
806
+ V.
807
+ DIFFERENT FP POTENTIALS OF THE
808
+ BMB LINE
809
+ We show in Fig. 7 the first derivative of different FP
810
+ potentials of the BMB line at N = ∞ and in d = 8/3.
811
+ These FP potentials, implicitly given by the exact ex-
812
+ pression given in Eqs. (9) and (10) of the main text, are
813
+ indexed by the nonnegative constant C.
814
+ The WF FP
815
+ potential corresponds to C = 0.
816
+ 1
817
+ 2
818
+ 3
819
+ 4
820
+ 5
821
+ 6
822
+ 7
823
+ μ
824
+ -0.20
825
+ -0.15
826
+ -0.10
827
+ -0.05
828
+ 0.05
829
+ 0.10
830
+ 0.15
831
+ V'[μ]
832
+ C=2
833
+ C=0.625
834
+ C=0
835
+ FIG. 7. ¯V ′(¯µ) for different FPs indexed by the constant C on
836
+ the BMB line given by Eqs. (9) and (10) of the main text.
837
+ We emphasize that the limit of ¯V ′′(¯µ0 = 3), when
838
+ C → 0 is not given by the second derivative of the WF
839
+ FP potential which is however the limit when C → 0 of
840
+ ¯V (¯µ) along the BMB line. This is consistent with what
841
+ happens at fixed d < 8/3 when N → ∞ since the limit
842
+ d → 8/3 at fixed α = ∞ consists in following the WF FP
843
+ at N = ∞ up to d = 8/3, the derivatives of which are
844
+ not the limit of the derivatives of the T3 potential.
845
+ VI.
846
+ EIGENPERTURBATIONS AT THE
847
+ TETRACRITICAL FP
848
+ FIG. 8.
849
+ d = 2.6: Eigenperturbation δ ¯V3(¯µ) at the T3 FP
850
+ corresponding, when N → ∞, to the relevant eigenvalue λ3 =
851
+ d − 2.
852
+ We show in Figs. 8 and 9 the relevant eigenperturba-
853
+ tions δ ¯Vi of the T3 FP in d = 2.6 for different values of
854
+ N. Whereas δ ¯V3 tends to the relevant eigenperturbation
855
+ of the critical WF FP –with eigenvalue d − 2 which is
856
+ the inverse of the critical exponent νWF–, the two others
857
+ become singular in the N → ∞ limit. This is the rea-
858
+ son why they play no role for the critical behavior of the
859
+ O(N) model at N = ∞.
860
+ 1
861
+ 2
862
+ 3
863
+ 4
864
+ 5
865
+ 6
866
+ 7 μ
867
+ -0.03
868
+ -0.02
869
+ -0.01
870
+ 0.00
871
+ 0.01
872
+ 0.02
873
+ 0.03
874
+ δV1[μ]
875
+ N=5×102
876
+ N=3×103
877
+ N=1×106
878
+ 0
879
+ 1
880
+ 2
881
+ 3
882
+ 4
883
+ 5
884
+ 6
885
+ 7 μ
886
+ 0.00
887
+ 0.01
888
+ 0.02
889
+ 0.03
890
+ 0.04
891
+ δV2[μ]
892
+ N=5×102
893
+ N=3×103
894
+ N=1.4×105
895
+ FIG. 9. d = 2.6: Eigenperturbations δ ¯Vn(¯µ) for n = 1, 2 at
896
+ the T3 FP corresponding respectively to the relevant eigen-
897
+ values λ1 ≃ 2.00 and λ2 ≃ 1.326 for different values of N.
898
+ These eigenperturbations tend to singular functions of ¯µ when
899
+ N → ∞.
900
+ We show in Fig. 10 that the slope of δ ¯V1(¯µ) at ¯µ0 in-
901
+ creases as N 1/3 which proves that this eigenperturbation
902
+ becomes discontinuous at infinite N.
903
+
904
+ [μ]
905
+ V[]
906
+ 0.05
907
+ 0.005
908
+ 0.04
909
+ 0.03
910
+ 0.000
911
+ 3.0
912
+ 35
913
+ 4.0
914
+ N=3x103
915
+ 0.02
916
+ 0.005
917
+ N=3x104
918
+ 0.01
919
+ WF N=8
920
+ 0.00
921
+ μ
922
+ 1
923
+ 2
924
+ 3
925
+ 5
926
+ 6
927
+ 0.01
928
+ -0.028
929
+ 1000
930
+ 104
931
+ 105
932
+ 106
933
+ N
934
+ 0.02
935
+ 0.05
936
+ 0.10
937
+ 0.20
938
+ -δV 1'[10/3]
939
+ FIG. 10. d = 2.6: Slope of the eigenperturbation δ ¯V1(¯µ) of
940
+ the T3 FP at its minimum ¯µ0 = 10/3 for different values of
941
+ N. The equation of the full line is y = 0.00175N 1/3.
942
+ VII.
943
+ BMB LINE AND THE JOINED LIMIT ϵ → 0
944
+ AND N → ∞ AT FIXED α = ϵN
945
+ When a T3 FP is followed along the hyperbola d =
946
+ 8/3 − α/N, α ≥ 0, of the (d, N) plane, its potential con-
947
+ verges when N → ∞ to the potential of one of the FPs
948
+ of the BMB line, see Fig. 11. We derive below the re-
949
+ lationship α = 162/C3 between the parameter α of the
950
+ hyperbola and the parameter C that indexes the FPs
951
+ along the BMB line, see Eqs. (9) and (10) of the main
952
+ text.
953
+ This relationship can be derived as follows. The FP
954
+ potential of T3 is expanded as
955
+ ¯V (¯µ) =
956
+
957
+
958
+ n=0
959
+ an(¯µ − 3)n
960
+ (1)
961
+ around the minimum ¯µ0 = 3 of the N = ∞ potential.
962
+ Then, the coefficients an are expanded as
963
+ an = a(0)
964
+ n
965
+ + N −1a(1)
966
+ n
967
+ + O(N −2)
968
+ (2)
969
+ in power of 1/N. At order O(N 0), Eq. (7) yields a(0)
970
+ n
971
+ = 0
972
+ for n = 1, 2 and 3 and recursively determines a(0)
973
+ n
974
+ for n
975
+ larger than 5 in terms of a(0)
976
+ 4 .
977
+ Now, �∞
978
+ n=0 a(0)
979
+ n (¯µ−3)n is the expansion of a FP poten-
980
+ tial of the BMB line. For this potential, ¯µ± behaves from
981
+ Eq. (9) as ¯µ± ≃ 3 ± 24/3C| ¯V ′|1/3 for C ̸= 0 and | ¯V ′| ≪ 1.
982
+ This implies that a(0)
983
+ 4
984
+ is related to C by a(0)
985
+ 4
986
+ = 3/(192C3).
987
+ At order O(N −1), it can be shown that a(1)
988
+ n
989
+ for all n but
990
+ n = 4 can be recursively determined in terms of a(0)
991
+ 4
992
+ or C,
993
+ if and only if the condition α = 162/C3 is satisfied which
994
+ proves the relationship between these two parameters.
995
+ We show in Fig. 11 different T3 FP potentials along
996
+ an hyperbola d = 8/3 − α/N with increasing values
997
+ of N.
998
+ These FP potentials converge to a potential
999
+ corresponding to the FP on the BMB line indexed by
1000
+ C = (162/α)1/3.
1001
+ 0
1002
+ 1
1003
+ 2
1004
+ 3
1005
+ 4
1006
+ 5
1007
+ 6
1008
+ 7
1009
+ μ
1010
+ -0.10
1011
+ -0.05
1012
+ 0.05
1013
+ 0.10
1014
+ V'[μ]
1015
+ C=0.672, N=∞
1016
+ N=4000 α=1600/3
1017
+ N=8000 α=1600/3
1018
+ N=24000 α=1600/3
1019
+ FIG. 11. ¯V ′(¯µ) of the T3 FP followed on the hyperbola d =
1020
+ 8/3 − α/N with fixed α = 1600/3 for increasing values of
1021
+ N. In the double limit d → 8/3 and N → ∞, it converges
1022
+ to the FP potential of the BMB line corresponding to C =
1023
+ (162 × 3/1600)1/3 ≃ 0.672.
1024
+
AtAzT4oBgHgl3EQfF_uW/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
AtFAT4oBgHgl3EQfrx4U/content/tmp_files/2301.08654v1.pdf.txt ADDED
@@ -0,0 +1,977 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Automated extraction of capacitive coupling for quantum dot systems
2
+ Joshua Ziegler,1, ∗ Florian Luthi,2 Mick Ramsey,2 Felix Borjans,2 Guoji Zheng,2 and Justyna P. Zwolak1, †
3
+ 1National Institute of Standards and Technology, Gaithersburg, MD 20899, USA
4
+ 2Intel Components Research, Intel Corporation, 2501 NW 229th Avenue, Hillsboro, Oregon 97124, USA
5
+ (Dated: January 23, 2023)
6
+ Gate-defined quantum dots (QDs) have appealing attributes as a quantum computing platform.
7
+ However, near-term devices possess a range of possible imperfections that need to be accounted
8
+ for during the tuning and operation of QD devices. One such problem is the capacitive cross-talk
9
+ between the metallic gates that define and control QD qubits. A way to compensate for the capacitive
10
+ cross-talk and enable targeted control of specific QDs independent of coupling is by the use of virtual
11
+ gates. Here, we demonstrate a reliable automated capacitive coupling identification method that
12
+ combines machine learning with traditional fitting to take advantage of the desirable properties of
13
+ each. We also show how the cross-capacitance measurement may be used for the identification of
14
+ spurious QDs sometimes formed during tuning experimental devices. Our systems can autonomously
15
+ flag devices with spurious dots near the operating regime which is crucial information for reliable
16
+ tuning to a regime suitable for qubit operations.
17
+ I.
18
+ INTRODUCTION
19
+ Quantum dot (QD) arrays, in which charge carriers
20
+ are trapped in localized potential wells and qubits can
21
+ be made by use of the spin and permutation symmetries
22
+ of the carriers, are a promising quantum computing plat-
23
+ form [1–3]. In fact, last year has shown the first demon-
24
+ stration of QD two-qubit gates with fidelities exceeding
25
+ the thresholds for fault-tolerant computing [4–6]. How-
26
+ ever, because the individual charge carriers that make
27
+ up qubits have electrochemical sensitivity to minor im-
28
+ purities and imperfections, calibration and tuning of QD
29
+ devices is a nontrivial and time-consuming process, with
30
+ each QD requiring a careful adjustment of a gate voltage
31
+ to define charge number, and multiple gate voltages to
32
+ specify tunnel coupling between QDs for two-qubit gates
33
+ or to reservoirs for reset and measurement. While manual
34
+ calibration is achievable for small, few-QD devices, with
35
+ increasing size and complexity of QD arrays, the relevant
36
+ control parameter space grows quickly, necessitating the
37
+ development of autonomous tuning methods.
38
+ There have been numerous demonstrations of automa-
39
+ tion of the various phases of the tuning process for sin-
40
+ gle and double-QD devices [7]. Some approaches seek to
41
+ tackle tuning starting from device turn-on to coarse tun-
42
+ ing [8–11] while others assume that bootstrapping (cal-
43
+ ibration of measurement devices and identification of a
44
+ nominal regime for further investigation) and basic tun-
45
+ ing (confirmation of controllability and device character-
46
+ istics) have been completed and focus on a more tar-
47
+ geted automation of the coarse and charge tuning [12–16].
48
+ While the initial auto-tuning approaches relied mainly
49
+ on the appealingly intuitive and relatively easy to imple-
50
+ ment conventional algorithms that typically involved a
51
+ ∗ Current address: Intel Components Research, Intel Corporation,
52
+ 2501 NW 229th Avenue, Hillsboro, Oregon 97124, USA
53
54
+ combination of techniques from regression analysis, pat-
55
+ tern matching, and quantum control theory, the more re-
56
+ cent algorithms take advantage of the modern computer
57
+ vision and machine learning [7].
58
+ A typical accumulation-mode QD device consists of
59
+ two sets of gates—plungers and barriers—that collec-
60
+ tively control the overall potential profile, QD-specific
61
+ single-particle energy detuning of individual QDs, the
62
+ tunnel couplings between QDs, and tunnel rates between
63
+ the most outer QDs and reservoirs. Ideally, each plunger
64
+ gate would affect only the electrochemical potential of
65
+ a single targeted QD and each barrier gate only one in-
66
+ tended tunnel barrier. Due to the tight proximity, how-
67
+ ever, each gate capacitively couples to nearby potential
68
+ and tunnel barriers. This makes careful control of these
69
+ key parameters challenging.
70
+ One way to compensate for the capacitive cross-talk
71
+ between gates is to enable orthogonal control of the QDs
72
+ potential by implementing so-called virtual gates [17].
73
+ Specifically, linear combinations of gate voltage changes
74
+ can be mapped onto onsite energy differences [17–20].
75
+ These approaches have been key for the initialization and
76
+ control of larger QD arrays [21, 22].
77
+ To autonomously identify capacitive couplings in a
78
+ device, various approaches have been demonstrated us-
79
+ ing both conventional fitting and machine learning (ML)
80
+ techniques [23–26]. However, these approaches, typically
81
+ relying on the Hough transform or conventional least-
82
+ squares fitting procedures, may be unreliable in the pres-
83
+ ence of data imperfections.
84
+ Hough transforms can ex-
85
+ tract slopes directly but may be sensitive to noise or be
86
+ excessively complex to analyze. The conventional fitting
87
+ can be more flexible but is susceptible to local minima
88
+ and can be time-consuming at inference time.
89
+ Convolutional neural networks (CNN) are well suited
90
+ for extracting high-level features from images and can
91
+ remain effective in the presence of noise or other imper-
92
+ fections [27]. However, ML methods can have difficulties
93
+ identifying data outside of the training distribution even
94
+ if it contains similar features [28]. Fortunately, given a
95
+ arXiv:2301.08654v1 [cond-mat.mes-hall] 20 Jan 2023
96
+
97
+ 2
98
+ simplified, high-level representation of the data, conven-
99
+ tional fitting approaches can be more targeted to extract
100
+ key information more effectively and quickly.
101
+ Here we develop a reliable automated capacitive cou-
102
+ pling identification method that combines ML with tra-
103
+ ditional fitting to take advantage of the desirable proper-
104
+ ties of each. We use an ML module for pixel classification
105
+ followed by linear regression for extracting targeted in-
106
+ formation and demonstrate effective performance across
107
+ noise levels and data variations. Testing each of these
108
+ methods on a set of eight simulated QD devices with
109
+ large variability and realistic noise variation mimicking
110
+ experimental conditions shows that the approach com-
111
+ bining ML and traditional fitting works well, with a root
112
+ mean square error (RMSE) of 0.034(14), corresponding
113
+ to a roughly 7 % error, for predicting virtual gate ma-
114
+ trix off-diagonal values (normalizing such that diagonal
115
+ values are one).
116
+ We also demonstrate how the cross-capacitance mea-
117
+ surement may be used for the identification of spurious
118
+ QDs formed during tuning experimental devices. Many
119
+ of the auto-tuning approaches proposed to date rely on
120
+ a series of small 2D scans capturing a relatively narrow
121
+ range of the voltage space [13, 14, 27, 29]. While such ap-
122
+ proaches improve the efficiency of tuning, they may result
123
+ in unexpected and difficult to assess failure modes when
124
+ the tuning algorithm terminates at an anti-crossing with
125
+ a spurious QD that may form in small potential wells due
126
+ to interface defects, surface roughness, or strain within
127
+ the device [30]. They are highly undesirable since they
128
+ may interfere with the QDs intended for use as qubits
129
+ and cannot themselves be used as qubits. To avoid de-
130
+ vice tuning failure, spurious QDs must be identified when
131
+ present and avoided. We test the utility of our approach
132
+ for capacitive coupling estimation by identifying spurious
133
+ QD in experimental measurements of QD devices [1].
134
+ The manuscript is organized as follows: In Sec. II we
135
+ introduce the framework of combining traditional fitting
136
+ techniques with a pixel classifier to process the high-level
137
+ information extracted from experimental data. In Sec. III
138
+ we show the utility of the proposed framework to auto-
139
+ matically extract virtual gates as well as identify charge
140
+ transitions resulting from a formation of spurious QD.
141
+ Finally, in Sec. IV we summarize the results and discuss
142
+ the outlook.
143
+ II.
144
+ METHODS: MACHINE LEARNING AND
145
+ FIT
146
+ Capacitive couplings in a QD device can be measured
147
+ and, in a constant capacitance approximation, described
148
+ by a matrix that maps the physical gate voltages onto
149
+ the effect they each have on the QD’s chemical poten-
150
+ tials or barriers [17, 23, 24, 31–33].
151
+ Measurement of
152
+ the elements of this matrix must be performed distinctly
153
+ for electrochemical potentials and tunnel barriers. Cou-
154
+ plings of the chemical potentials to each QD—which is
155
+ 0.30
156
+ 0.32
157
+ 0.27
158
+ 0.29
159
+ VP2(V)
160
+ (a)
161
+ 0.8
162
+ 1.1
163
+ 1.4
164
+ Current (arb. units)
165
+ 0.30
166
+ 0.32 VP1(V)
167
+ (b)
168
+ NT LT CT RT PL
169
+ (c)
170
+ FIG. 1. An example 2D scan and corresponding pixel classifi-
171
+ cation, class clusters, and linear fits. (a) A simulated voltage
172
+ scan showing left and right transitions as well as a polariza-
173
+ tion line. (b) Pixel classification for the scan shown in (a). (c)
174
+ Regions of pixels and linear fits from the pixel classification.
175
+ The large dark points indicate the centers of pixel regions.
176
+ the focus of this work—can be extracted from shifts in
177
+ charge transition lines when each voltage is varied [17]
178
+ while the effect of each gate on tunnel barriers can be
179
+ assessed by measuring changes in the width of inter-dot
180
+ transitions, assuming the electron temperature is suffi-
181
+ ciently low [32]. Measured this way, the couplings are
182
+ relative, usually scaled with respect to the coupling of
183
+ the QD to the nearest gate. An absolute energy scale
184
+ can be obtained by measuring the gate lever arms with
185
+ photon-assisted tunneling, Coulomb diamonds, or bias
186
+ triangles [34]. However, for establishing the orthogonal
187
+ control the relative scale is sufficient [21].
188
+ For a double QD, the virtualization matrix Gvirt re-
189
+ lating the physical plunger gates to virtual gates can be
190
+ represented by Eq. 1. Each row is normalized such that
191
+ the diagonal entries are 1 to reflect the relative nature of
192
+ our virtual gates.
193
+ Gvirt ≡
194
+ �VP ′
195
+ 1
196
+ VP ′
197
+ 2
198
+
199
+ =
200
+
201
+ 1
202
+ α12
203
+ α21
204
+ 1
205
+ � �
206
+ VP1
207
+ VP2
208
+
209
+ (1)
210
+ The relative cross-capacitances for chemical potentials
211
+ manifest themselves via the slopes of charge transition
212
+ lines, with the dominant terms of the cross-capacitance
213
+ matrix determined from a measurement in the space of
214
+ neighboring pairs of gates [21].
215
+ We address the iden-
216
+ tification of the cross-capacitances as captured in two-
217
+ dimensional (2D) plunger-plunger gate scans, as shown in
218
+ Fig. 1(a). To translate the low-level QD data into high-
219
+ level information useful for automation we use a pixel
220
+ classifier, i.e., a CNN model with a structure similar to a
221
+ feature pyramid network [35]. The pixel classifier takes as
222
+
223
+ 3
224
+ an input a small 2D plunger voltage scan obtained using
225
+ a charge sensor, as shown in Fig. 1(a). It then identi-
226
+ fies each pixel within the scan as belonging to one of the
227
+ charge transition classes—left, right, central, or inter-dot
228
+ (polarization line) transition, denoted as LT, RT, CT, or
229
+ PL, respectively—or to the no transition (NT) class. In
230
+ other words, the CNN provides a high-level classification
231
+ of the raw experimental data while keeping spatial infor-
232
+ mation about the relative location and orientation of the
233
+ detected features, which is useful for algorithmic process-
234
+ ing. Figure 1(b) shows the pixel classification of a scan
235
+ from Fig. 1(a).
236
+ To translate pixel classifications to capacitive cou-
237
+ plings, we identify contiguous regions within each class
238
+ of pixels in an image. A labeling algorithm from the mul-
239
+ tidimensional image processing package in SciPy is then
240
+ used to determine the relevant clusters of connected pix-
241
+ els [36]. This separates fragments of charge transitions
242
+ into distinct clusters so that each can be processed indi-
243
+ vidually. Each region of pixels classified as LT, CT, or
244
+ RT is then independently fitted using linear regression, as
245
+ shown in Fig. 1(c). When multiple segments for a given
246
+ class are present in an image, the capacitive coupling re-
247
+ turned is the average for all fitted lines weighted by the
248
+ standard deviations of each fit, yielding the solid lines in
249
+ Fig. 1(b) (offset arbitrarily for comparison with the pixel
250
+ regions). Standard deviations σ are computed from the
251
+ standard error of the fit, S, by σ = S/√n, where n is the
252
+ size of the pixel region, as in Student’s t-distribution [37].
253
+ In addition, each region is tagged with its center in volt-
254
+ age space, shown by the large black points in Fig. 1(c),
255
+ which allows tracking the changes in charge transitions
256
+ and their slopes within the larger space.
257
+ A.
258
+ Data
259
+ The data used for training the ML tools and testing
260
+ the methods was generated using a simulation of QD de-
261
+ vices [12]. The simulation is composed of a calculation of
262
+ the electron density in the Thomas-Fermi approximation
263
+ and a capacitance matrix to determine the stable electron
264
+ configuration. To improve the robustness of the models,
265
+ the data is augmented with synthetic white, pink (1/f),
266
+ and telegraph noise [27]. The effect of a QD charge sen-
267
+ sor strongly coupled to the plunger gates is varied during
268
+ the scan to improve performance on this type of experi-
269
+ mental data.
270
+ The training dataset consists of 1.6 × 105 devices with
271
+ parameters varied over a uniform distribution with a
272
+ standard deviation equal to 1 % of each parameter’s
273
+ value. To train the ML models we randomly sample 10
274
+ small scans per device and use charge state ground truth
275
+ to label each scan on a pixel level with the presence and
276
+ type of transition, yielding NT, LT, CT, RT, and PL la-
277
+ bels. Additionally, we extract the slopes of the transition
278
+ lines directly using the gradients of the device charge.
279
+ The test data is composed of eight simulated devices
280
+ with large variations in screening length and device pitch
281
+
282
+
283
+ 10×
284
+ 15×
285
+ 20×
286
+ 25×
287
+ 30×
288
+ 35×
289
+ 0.0
290
+ 0.2
291
+ 0.4
292
+ 0.6
293
+ RMSE
294
+ (a)
295
+
296
+ 15×
297
+ 30×
298
+ 0.0
299
+ 0.2
300
+ 0.4
301
+ 0.6
302
+ (b)
303
+
304
+ 15×
305
+ 30×
306
+ Noise Level
307
+ (c)
308
+ FIG. 2. (a) Root mean square error (RMSE) for all transition
309
+ classes (left, central, and right [LT, CT, RT]) as a function
310
+ of the synthetic noise level. (b) RMSE as a function of noise
311
+ level for the LT class. (c) RMSE as a function of noise level
312
+ for the RT class.
313
+ and with large shifts in the position of one of the plunger
314
+ gates. These changes lead to large variations in the slopes
315
+ of the charge transition lines, the capacitive coupling
316
+ between QDs, spacing between lines, and the relative
317
+ amount of left and right QD, making them largely dis-
318
+ tinct from the training data. To facilitate a controlled
319
+ study and track the performance of the pixel classifier
320
+ as data quality degrades, each large scan is randomly
321
+ sampled 50 times and the resulting small scans are aug-
322
+ mented with increasing levels of synthetic noise.
323
+ This
324
+ results in a set of 400 simulated test scans. Finally, sev-
325
+ eral large experimental measurements acquired using a
326
+ double-QD configuration on a three-QD Six/SiGe1−x de-
327
+ vice, fabricated on an industrial 300 mm process line [1],
328
+ are used to test the performance of the virtualization
329
+ algorithm. Experimental scans capturing spurious QDs
330
+ are used to demonstrate the algorithm for spurious QD
331
+ detection.
332
+ III.
333
+ RESULTS
334
+ We test the effectiveness of our automated approach to
335
+ extracting the cross-capacitance by first evaluating the
336
+ performance of each component, i.e., the pixel classifier
337
+ and the slope extractions, on each scan in the simulated
338
+ test set. The error of the pixel classifier in our frame-
339
+ work is defined as a fraction of pixels belonging to true
340
+ transitions that are not contained in line segments in the
341
+ CNN output.
342
+ This captures type-I errors without the
343
+
344
+ 4
345
+ 2.00
346
+ 2.20
347
+ VP1(V)
348
+ 2.20
349
+ 2.40
350
+ VP2(V)
351
+ (a)
352
+ 1.40
353
+ 1.55 VP ′
354
+ 1(V)
355
+ 1.50
356
+ 1.60
357
+ 1.70
358
+ VP ′
359
+ 2(V)
360
+ (b)
361
+ 1.20
362
+ 1.35 VP ′′
363
+ 1 (V)
364
+ 1.20
365
+ 1.30
366
+ VP ′′
367
+ 2 (V)
368
+ (c)
369
+ −0.8
370
+ −0.6
371
+ −0.4
372
+ −0.2
373
+ Virt. gate
374
+ off diag.
375
+ FIG. 3. (a) Large experimentally measured charge stability diagram with a scatter plot of centers of pixel class regions overlaid.
376
+ The colors of the points indicate the virtual gate off-diagonal values identified by fits to the region. The sizes of the points
377
+ indicate the weights used when averaging. Only points with relative error less than 20 % are plotted. (a,b) Charge stability
378
+ diagram after applying virtual gates acquired near the (0, 0) − (1, 1) charge transition in (a) and near the (1, 3) − (2, 4) charge
379
+ transition in (b). In both (b) and (c) the virtualization is performed off-line, via an affine transform to the original scan shown
380
+ in (a) and the points are plotted using the same parameters as in (a).
381
+ effect of false type-II errors due to imperfect labels [38].
382
+ Figure 2(a) shows the change in RMSE as a function of
383
+ the noise level in the simulated data. At the noise level of
384
+ 1.0, i.e., the noise level estimated from experimental data
385
+ in Ref. [29], we observe an RMSE of 0.17(5). The RMSE
386
+ increases significantly to 0.50(11) at the noise level of 20.
387
+ For reference, a pixel classifier that always predicts the
388
+ NT class would have an RMSE of 0.62 (
389
+
390
+ 0.4). For the
391
+ LT and RT classes relevant to cross-capacitances, shown
392
+ in Fig. 2(b) and (c), the pixel classifier for noise level 1.0
393
+ has an RMSE of 0.20(8) and 0.11(8), respectively.
394
+ To verify that the slope extraction tool works as in-
395
+ tended, we test it across the eight large simulated test
396
+ devices. For these tests, we evaluate the pixel classifier
397
+ in windows of size roughly 1.5× the charging energy, as
398
+ estimated by the spacing of the first two charge transi-
399
+ tions. Outputs from the pixel classifier are cropped by
400
+ one pixel from the edge of the image before processing
401
+ due to missing context reducing CNN performance [39].
402
+ The resulting classes of pixels are then grouped into dis-
403
+ tinct clusters. For each cluster consisting of more than
404
+ five pixels an independent linear fit is performed, return-
405
+ ing both the slope and the standard error of the fitted
406
+ line. This information can be used to find the orthog-
407
+ onal “virtual” control space or to flag transitions that
408
+ potentially belong to spurious dots, as described in the
409
+ following sections.
410
+ A.
411
+ Deriving virtual gates
412
+ As stated in Sec. II, in our approach the off-diagonal
413
+ elements, defining the virtual gates transformation, are
414
+ determined based on the slopes of the LT and RT cap-
415
+ tured in a given image, and the diagonal elements of the
416
+ capacitance matrix are set to 1.0. When multiple lines
417
+ belonging to the same class are detected, as in Fig. 1(a),
418
+ the capacitive coupling is calculated through a weighted
419
+ average, with the weight accounting for both the size
420
+ of the clusters and the standard deviations of respective
421
+ fits [37].
422
+ The off-diagonal elements of the virtualization matrix
423
+ computed this way have an RMSE of 0.034(14) at the
424
+ noise level of 1.0 defined in Ref. [29], corresponding to a
425
+ roughly 8 % error compared to the ground truth values
426
+ derived from simulated data. We further test them on
427
+ a range of levels of synthetic noise and find the RMSE
428
+ rises by a factor of two at a level of noise of roughly 15×
429
+ the level of noise defined in Ref. [29], consistent with the
430
+ pixel classifier error.
431
+ To better understand the trends of the virtualization
432
+ matrix in the plunger-plunger space, we carry out a per-
433
+ formance analysis using the test set of eight large sim-
434
+ ulated charge stability diagrams and several experimen-
435
+ tally measured scans.
436
+ For each scan, we calculate the
437
+ fits to the pixel classification clusters based on a series of
438
+ small scans sampled at each point within the large scan
439
+ with the exclusion of a margin implemented to ensure
440
+ that all sampled scans fall within the full scan bound-
441
+ aries. The small scans and the margins are set to have a
442
+ size 1.5× the charging energy of a given simulated device.
443
+ Figure 3(a) shows the centers of the pixel region identi-
444
+ fied in each small scan [as in Fig. 1(c)] as the sampling
445
+ window is swept across a large experimentally measured
446
+ charge stability diagram. The regions identified by the
447
+ pixel classification are consistently placed correctly on
448
+ the charge transition lines regardless of the position of the
449
+
450
+ 5
451
+ 2.09
452
+ 2.18 VP1(V)
453
+ 0.00
454
+ 0.25
455
+ 0.50
456
+ Rel. virt. gate off-diag.
457
+ (a)
458
+ 2.26
459
+ 2.34 VP2(V)
460
+ (b)
461
+ 0
462
+ 1
463
+ 2
464
+ 3
465
+ 4
466
+ Density
467
+ ×10−2
468
+ FIG. 4. Histograms of the off-diagonal elements of the virtu-
469
+ alization matrix for an experimentally measured scan shown
470
+ in Fig. 3(a) as a function of plunger gates, (a) VP1 and (b)
471
+ VP2. Off-diagonal values are normalized to the mean of the
472
+ virtual gates in the (1, 1) charge state for ease of comparison.
473
+ Virtual gate values are extracted from a strip of small scans
474
+ shifted by 6 mV (four pixels) to better visualize variation at
475
+ each plunger gate value.
476
+ line within a small scan. Region centers shift along the
477
+ charge transition lines as different portions of the line are
478
+ captured within the small scan and remain fixed when-
479
+ ever the same fragment of the charge transition is cap-
480
+ tured. The color of the points indicates the off-diagonal
481
+ values of the virtual gate matrix, α12 and α21. As ex-
482
+ pected, these coupling constants get larger in magnitude
483
+ as charges are added to each QD. Finally, the size of the
484
+ points in Fig. 3(a) indicates the 1/σ2 weight of the slopes
485
+ used when averaging multiple slopes from the same type
486
+ of transition within a small scan. As desired, the posi-
487
+ tions of the points with smaller sizes indicate that lines
488
+ that are smaller or less captured within a small scan have
489
+ fits with larger errors. Overall, this plot confirms that the
490
+ pixel classification and the fits are working as intended
491
+ at capturing charge transition lines and their slopes.
492
+ To demonstrate the spatial relevance of the virtual
493
+ gates derived from a set of fits across a device’s charge
494
+ landscape, in Fig. 3(b) and (c) we plot affine-transformed
495
+ charge stability diagrams, with points indicating fits
496
+ overlaid. The points plotted are the centers of pixel re-
497
+ gions with colors indicating the α12 and α21 values and
498
+ size indicating the inverse of the fit error squared (the
499
+ weight of the fit in the average).
500
+ The affine transfor-
501
+ mation applied in Fig. 3(b) corresponds to virtual gates
502
+ derived from an image near the (0, 0) − (1, 1) charge
503
+ transition with off-diagonal values α12 = −0.282(4),
504
+ α21 = −0.331(4). For Fig. 3(c), the affine transformation
505
+ applied has virtual gates from the (1, 3) − (2, 4) charge
506
+ transition, with off-diagonal values α12 = −0.363(4),
507
+ α21 = −0.480(4). As can be seen in the insets in Fig. 3(b)
508
+ and (c), these virtual gates are very effective at trans-
509
+ forming the target region to an orthogonal space, but the
510
+ difference between the extracted virtual gate off-diagonal
511
+ values are about 50 % higher for the latter case. This
512
+ highlights the importance of an efficient local method for
513
+ determining virtual gates.
514
+ To further understand how capacitive coupling varies
515
+ across a charge stability diagram, we can calculate varia-
516
+ tion as each plunger gate is adjusted. Figure 4(a) and (b)
517
+ show how virtual gates extracted from small scans change
518
+ as VP1 and VP2 are varied. To better show the trend, vir-
519
+ tual gates from small scans shifted by 3 mV (two pixels)
520
+ in the opposing direction are included. This shows that
521
+ the virtual gates extracted from small scans effectively
522
+ capture variation across charge stability diagrams.
523
+ B.
524
+ Detection of spurious dots
525
+ Visually, spurious QDs are recognized in large 2D scans
526
+ as charge transitions with slopes diverging from a mono-
527
+ tonic trend, see Fig. 5(a). In this framework, they may
528
+ be identified as transition lines with anomalous capaci-
529
+ tive couplings relative to the transitions around them.
530
+ As a demonstration, we use the pixel classification and
531
+ fit tools to analyze five experimental charge stability di-
532
+ agrams: two capturing properly formed QD, shown in
533
+ Fig. 5(a) and (b), and three capturing spurious QDs,
534
+ shown in Fig. 5(a), (b), and (c). While for extraction
535
+ of the virtualization matrix small scans are sufficient, de-
536
+ tection of spurious QD requires somewhat bigger scans
537
+ to ensure that the neighboring charge transitions are ad-
538
+ equately captured. In our analysis, we rely on 2D scans
539
+ of a size roughly three times the charging energy (four
540
+ times the area of scans typically used in auto-tuning al-
541
+ gorithms [13, 14]). We also consider only clusters con-
542
+ sisting of at least 20 pixels to ensure better reliability of
543
+ the linear fit.
544
+ After pixel classification, contiguous clusters of pixels
545
+ belonging to a given class of transitions are analyzed in-
546
+ dividually, resulting in a cluster-based fit and standard
547
+ deviation. Cases where more than one cluster belongs
548
+ to a given charge transition result in separate fits, as in
549
+ Fig. 5(b) and (e) where the LT lines are split into groups
550
+ to either side of the RT lines. This separation serves two
551
+ purposes: to ensure that variation along a given transi-
552
+ tion isn’t included and to treat each additional line inde-
553
+ pendent of the charge on another QD.
554
+ Within a class and group of transitions, the magni-
555
+ tude of the capacitive coupling is expected to increase
556
+ monotonically as the charge is added. Such behavior is
557
+ clearly visible in Fig. 5(k) and (l), with the latter having
558
+ to separate groups of fits (shown with different shades of
559
+ green) for the groups of clusters. On the contrary, a spu-
560
+ rious QD manifests itself by a non-monotonic behavior of
561
+ the capacitive coupling between transitions, as depicted
562
+ graphically by the center point (or group of points) in
563
+ Fig. 5(m), (n), and (o). The severity of this divergence
564
+ can be quantified using a Z-test [40].
565
+ In practical applications, the automated detection of
566
+ spurious QD fits nicely within the auto-tuning paradigm.
567
+ As mentioned earlier, many of the proposed approaches
568
+ utilize a series of small 2D scans [13, 14, 27, 29] or
569
+ 1D rays [41, 42] as means to improve the tuning effi-
570
+
571
+ 6
572
+ 2.26
573
+ 2.32
574
+ 2.33
575
+ 2.39
576
+ VP2(V)
577
+ (a)
578
+ 2.26
579
+ 2.32
580
+ 2.33
581
+ 2.39
582
+ (f)
583
+ 2.33
584
+ 2.39
585
+ -0.31
586
+ -0.28
587
+ -0.26
588
+ α12
589
+ (k)
590
+ 2.38
591
+ 2.44
592
+ 2.31
593
+ 2.37
594
+ VR1
595
+ (b)
596
+ 0.6
597
+ 0.8
598
+ 1.0
599
+ 1.2
600
+ Current
601
+ (arb. units)
602
+ 2.38
603
+ 2.44
604
+ 2.31
605
+ 2.37
606
+ (g)
607
+ NT
608
+ LT
609
+ CT
610
+ RT
611
+ PL
612
+ 2.31
613
+ 2.37
614
+ -0.4
615
+ -0.3
616
+ -0.2 (l)
617
+ 1.89
618
+ 1.96
619
+ 2.18
620
+ 2.24
621
+ (c)
622
+ 1.89
623
+ 1.94
624
+ 2.18
625
+ 2.24
626
+ (h)
627
+ 2.18
628
+ 2.25
629
+ -0.4
630
+ -0.3
631
+ -0.2 (m)
632
+ 0.42
633
+ 0.46
634
+ 0.46
635
+ 0.50
636
+ (d)
637
+ 0.42
638
+ 0.46
639
+ 0.46
640
+ 0.50
641
+ (i)
642
+ 0.46
643
+ 0.50
644
+ -0.5
645
+ -0.4
646
+ -0.3 (n)
647
+ 0.48
648
+ 0.52
649
+ 0.45
650
+ 0.49
651
+ VR2
652
+ (e)
653
+ 0.7
654
+ 0.8
655
+ 0.9
656
+ 1.0
657
+ Current
658
+ (arb. units)
659
+ 0.48
660
+ 0.52 VP1(V)
661
+ 0.45
662
+ 0.49
663
+ (j)
664
+ NT
665
+ LT
666
+ CT
667
+ RT
668
+ PL
669
+ 0.45
670
+ 0.49 VP2(V)
671
+ -0.6
672
+ -0.4
673
+ -0.2 (o)
674
+ FIG. 5. Spurious dot detection. The top row shows two charge stability diagrams capturing properly formed QD [panels (a)
675
+ and (b)] and three charge stability capturing spurious QD [panels (c), (d), and (e)]. The black boxes in (b) and (e) in highlight
676
+ small 2D scans, denoted as VR1 and VR2, typical of the auto-tuning approaches proposed in Ref. [13, 14]. Panels (f)–(j) in the
677
+ middle row show pixel classification results for charge stability diagrams shown in the top row. Plots of fitting results used to
678
+ determine whether a spurious QD is present are shown in the bottom row [panels (k)–(o)]. The different groups of transition
679
+ are shown with different shades of green. The monotonicity within each group of transitions is clearly visible in panels (k) and
680
+ (l). On the contrary, in the three plots shown in panels (m), (n), and (o), there is a clear divergence from the expected trend
681
+ for the spurious QD (middle) transitions. Error bars indicate one standard deviation.
682
+ ciency.
683
+ While these approaches deliver measurement-
684
+ cost-effective solutions, they are prone to unexpected and
685
+ difficult-to-detect failure even when the data quality is
686
+ high. Fig. 5(b) and (e) show examples of such potentially
687
+ problematic cases. The small 2D regions in the plunger-
688
+ plunger space, highlighted in these scans with the black
689
+ boxes, are typical for topology setting algorithms.
690
+ In
691
+ both cases, they are classified by a state classifier model
692
+ as double QD state, with state prediction vectors being
693
+ p(VR1) = [0.01, 0.04, 0., 0.04, 0.92] for VR1 region high-
694
+ lighted in Fig. 5(b) and p(VR2) = [0., 0., 0.18, .05, 0.76]
695
+ for region VR2 highlighted in Fig. 5(b), where p(VR) =
696
+ [pND, pSDL, pSDC, pSDR, pDD] with ND denoting no QDs
697
+ formed, SDL, SDC, and SDR denoting the left, central,
698
+ and right single QD, respectively, and DD denoting the
699
+ double-QD state. Moreover, the data quality for these
700
+ images is high in both cases, with q(VR1) = [1.0, 0.0, 0.0]
701
+ for region VR1 and q(VR1) = [0.99, 0.01, 0.0] for VR2,
702
+ where q(VR) = [phigh, pmod, plow] with phigh, pmod, and
703
+ plow denoting the probability of region VR being assessed
704
+ by the data quality control module as “high,”, “moder-
705
+ ate,” and “low” quality, respectively. Thus, from the ML
706
+ perspective, both these predictions are confidently cor-
707
+ rect. However, when looked at within a slightly larger
708
+ voltage range, it is clear that in the latter case the
709
+ small scan captures an anti-crossing with a spurious QD,
710
+ which for practical tuning purposes is a failure. If not
711
+ recognized and corrected for, termination at this point
712
+ will result in an incorrect charge setting and virtualiza-
713
+ tion [13, 29].
714
+ The spurious QD detection algorithm can be easily
715
+ implemented in the auto-tuning algorithm proposed in
716
+ Ref. [27] as a safety check before the unloading step is ini-
717
+ tiated. An automated identification and characterization
718
+ of spurious QDs may also be useful to inform fabrication
719
+ procedures and prevent them in future devices [30].
720
+ IV.
721
+ CONCLUSIONS
722
+ As quantum dot devices grow in size and complexity,
723
+ the need for reliable and automated tune-up procedures
724
+ becomes more pressing.
725
+ Establishing orthogonal con-
726
+ trol of the chemical potentials of quantum dots is one
727
+ of the first steps in the tune-up of any larger quantum
728
+ dot array. Here, we demonstrated a method that com-
729
+ bines machine-learning-based pixel-classification and tra-
730
+ ditional curve fitting to reliably determine voltage cross-
731
+ talk coefficients. The advantage of this method over pre-
732
+ vious approaches is highlighted by increased reliability
733
+
734
+ 7
735
+ and resilience to experimental noise.
736
+ Further on, un-
737
+ wanted spurious dots that would reduce or inhibit de-
738
+ vice performance can be detected and flagged when this
739
+ module is used as part of a larger tune-up algorithm [29].
740
+ The capability to automatically and reliably detect spuri-
741
+ ous dots is especially important on wafer-scale fabrication
742
+ characterization tools that produce more data than can
743
+ efficiently be processed by human analysis. In extensions,
744
+ our tools could allow for automated navigation of volt-
745
+ age space for more targeted measurement of all chemical
746
+ potential and tunnel barrier cross-capacitances [17, 32].
747
+ ACKNOWLEDGMENTS
748
+ This research was performed while J.Z. held a NRC
749
+ Research Associateship award at the National Institute
750
+ of Standards and Technology (NIST). The views and con-
751
+ clusions contained in this paper are those of the authors
752
+ and should not be interpreted as representing the of-
753
+ ficial policies, either expressed or implied, of the U.S.
754
+ Government.
755
+ The U.S. Government is authorized to
756
+ reproduce and distribute reprints for Government pur-
757
+ poses notwithstanding any copyright noted herein. Any
758
+ mention of commercial products is for information only;
759
+ it does not imply recommendation or endorsement by
760
+ NIST.
761
+ [1] R. Pillarisetty,
762
+ T. Watson,
763
+ B. Mueller,
764
+ E. Henry,
765
+ H. George, S. Bojarski, L. Lampert, F. Luthi, R. Kotl-
766
+ yar, O. Zietz, S. Neyens, F. Borjans, R. Caudillo,
767
+ D. Michalak, R. Nahm, J. Park, M. Ramsey, J. Roberts,
768
+ S. Schaal, G. Zheng, T. Kr¨ahenmann, M. Lodari, A. Zw-
769
+ erver, M. Veldhorst, G. Scappucci, L. Vandersvpen, and
770
+ J. Clarke, Si mos and si/sige quantum well spin qubit
771
+ platforms for scalable quantum computing, in 2021 IEEE
772
+ International Electron Devices Meeting (IEDM) (2021)
773
+ pp. 14.1.1–14.1.4.
774
+ [2] P. L. Bavdaz, H. G. J. Eenink, J. van Staveren, M. Lo-
775
+ dari, C. G. Almudever, J. S. Clarke, F. Sebasatiano,
776
+ M. Veldhorst, and G. Scappucci, A quantum dot crossbar
777
+ array with sublinear scaling of interconnects at cryogenic
778
+ temperature, npj Quantum Information 8, 1 (2022).
779
+ [3] J. M. Boter, J. P. Dehollain, J. P. van Dijk, Y. Xu,
780
+ T. Hensgens, R. Versluis, H. W. Naus, J. S. Clarke,
781
+ M. Veldhorst, F. Sebastiano, and L. M. Vandersypen,
782
+ Spiderweb array: A sparse spin-qubit array, Phys. Rev.
783
+ Applied 18, 024053 (2022).
784
+ [4] A. Noiri, K. Takeda, T. Nakajima, T. Kobayashi, A. Sam-
785
+ mak, G. Scappucci, and S. Tarucha, A shuttling-based
786
+ two-qubit logic gate for linking distant silicon quantum
787
+ processors, Nat. Commun. 13, 5740 (2022).
788
+ [5] X. Xue, M. Russ, N. Samkharadze, B. Undseth, A. Sam-
789
+ mak, G. Scappucci, and L. M. K. Vandersypen, Quantum
790
+ logic with spin qubits crossing the surface code threshold,
791
+ Nature 601, 343 (2022).
792
+ [6] A. R. Mills, C. R. Guinn, M. J. Gullans, A. J. Sigillito,
793
+ M. M. Feldman, E. Nielsen, and J. R. Petta, Two-qubit
794
+ silicon quantum processor with operation fidelity exceed-
795
+ ing 99%, Sci. Adv. 8, eabn5130 (2022).
796
+ [7] J. P. Zwolak and J. M. Taylor, Colloquium:
797
+ Ad-
798
+ vances in automation of quantum dot devices control,
799
+ arXiv:2112.09362 10.48550/arxiv.2112.09362 (2021).
800
+ [8] T. A. Baart, P. T. Eendebak, C. Reichl, W. Wegscheider,
801
+ and L. M. K. Vandersypen, Computer-automated tuning
802
+ of semiconductor double quantum dots into the single-
803
+ electron regime, Appl. Phys. Lett. 108, 213104 (2016).
804
+ [9] J. Darulov´a, S. J. Pauka, N. Wiebe, K. W. Chan, G. C.
805
+ Gardener, M. J. Manfra, M. C. Cassidy, and M. Troyer,
806
+ Autonomous tuning and charge-state detection of gate-
807
+ defined quantum dots, Phys. Rev. Applied 13, 054005
808
+ (2020).
809
+ [10] H. Moon, D. T. Lennon, J. Kirkpatrick, N. M. van Es-
810
+ broeck, L. C. Camenzind, L. Yu, F. Vigneau, D. M.
811
+ Zumb¨uhl, G. A. D. Briggs, M. A. Osborne, D. Sejdinovic,
812
+ E. A. Laird, and N. Ares, Machine learning enables com-
813
+ pletely automatic tuning of a quantum device faster than
814
+ human experts, Nat. Commun. 11, 4161 (2020).
815
+ [11] S. Czischek, V. Yon, M.-A. Genest, M.-A. Roux, S. Ro-
816
+ chette, J. C. Lemyre, M. Moras, M. Pioro-Ladri´ere,
817
+ D. Drouin, Y. Beilliard, and R. G. Melko, Miniaturizing
818
+ neural networks for charge state autotuning in quantum
819
+ dots, Mach. Learn.: Sci. Technol. 3, 015001 (2021).
820
+ [12] J. P. Zwolak, S. S. Kalantre, X. Wu, S. Ragole, and J. M.
821
+ Taylor, QFlow lite dataset: A machine-learning approach
822
+ to the charge states in quantum dot experiments, PLoS
823
+ ONE 13, 1 (2018).
824
+ [13] R. Durrer, B. Kratochwil, J. Koski, A. Landig, C. Reichl,
825
+ W. Wegscheider, T. Ihn, and E. Greplova, Automated
826
+ tuning of double quantum dots into specific charge states
827
+ using neural networks, Phys. Rev. Applied 13, 054019
828
+ (2020).
829
+ [14] J. P. Zwolak, T. McJunkin, S. S. Kalantre, J. Dodson,
830
+ E. MacQuarrie, D. Savage, M. Lagally, S. Coppersmith,
831
+ M. A. Eriksson, and J. M. Taylor, Autotuning of double-
832
+ dot devices in situ with machine learning, Phys. Rev.
833
+ Applied 13, 034075 (2020).
834
+ [15] J. P. Zwolak, T. McJunkin, S. S. Kalantre, S. F. Neyens,
835
+ E. R. MacQuarrie, M. A. Eriksson, and J. M. Taylor,
836
+ Ray-based framework for state identification in quantum
837
+ dot devices, PRX Quantum 2, 020335 (2021).
838
+ [16] M. Lapointe-Major, O. Germain, J. Camirand Lemyre,
839
+ D. Lachance-Quirion, S. Rochette, F. Camirand Lemyre,
840
+ and M. Pioro-Ladri`ere, Algorithm for automated tuning
841
+ of a quantum dot into the single-electron regime, Phys.
842
+ Rev. B 102, 085301 (2020).
843
+ [17] T. Hensgens,
844
+ T. Fujita,
845
+ L. Janssen,
846
+ X. Li,
847
+ C. J.
848
+ Van Diepen, C. Reichl, W. Wegscheider, S. D. Sarma,
849
+ and L. M. K. Vandersypen, Quantum simulation of a
850
+ fermi–hubbard model using a semiconductor quantum
851
+
852
+ 8
853
+ dot array, Nature 548, 70 (2017).
854
+ [18] T. H. Oosterkamp, T. Fujisawa, W. G. van der Wiel,
855
+ K. Ishibashi, R. V. Hijman, S. Tarucha, and L. P.
856
+ Kouwenhoven, Microwave spectroscopy of a quantum-dot
857
+ molecule, Nature 395, 873 (1998).
858
+ [19] T. Hensgens, Emulating Fermi-Hubbard physics with
859
+ quantum dots: from few to more and how to, Ph.D. the-
860
+ sis, Delft University of Technology, Delft, Netherlands
861
+ (2018).
862
+ [20] J. K. Perron, M. D. Stewart Jr, and N. M. Zimmerman, A
863
+ quantitative study of bias triangles presented in chemical
864
+ potential space, J. Phys.: Condens. Matter 27, 235302
865
+ (2015).
866
+ [21] C. Volk, A. M. J. Zwerver, U. Mukhopadhyay, P. T. Een-
867
+ debak, C. J. van Diepen, J. P. Dehollain, T. Hensgens,
868
+ T. Fujita, C. Reichl, W. Wegscheider, and L. M. K. Van-
869
+ dersypen, Loading a quantum-dot based “Qubyte” regis-
870
+ ter, npj Quantum Inf. 5, 29 (2019).
871
+ [22] A. R. Mills, D. M. Zajac, M. J. Gullans, F. J. Schupp,
872
+ T. M. Hazard, and J. R. Petta, Shuttling a single charge
873
+ across a one-dimensional array of silicon quantum dots,
874
+ Nat. Commun. 10, 1063 (2019).
875
+ [23] C. J. van Diepen, P. T. Eendebak, B. T. Buijtendorp,
876
+ U. Mukhopadhyay, T. Fujita, C. Reichl, W. Wegscheider,
877
+ and L. M. K. Vandersypen, Automated tuning of inter-
878
+ dot tunnel coupling in double quantum dots, Appl. Phys.
879
+ Lett. 113, 033101 (2018).
880
+ [24] A. R. Mills, M. M. Feldman, C. Monical, P. J. Lewis,
881
+ K. W. Larson, A. M. Mounce, and J. R. Petta, Computer-
882
+ automated tuning procedures for semiconductor quan-
883
+ tum dot arrays, Appl. Phys. Lett. 115, 113501 (2019).
884
+ [25] G. A. Oakes, J. Duan, J. J. L. Morton, A. Lee, C. G.
885
+ Smith, and M. F. G. Zalba, Automatic virtual voltage
886
+ extraction of a 2x2 array of quantum dots with machine
887
+ learning (2020).
888
+ [26] H. Liu, B. Wang, N. Wang, Z. Sun, H. Yin, H. Li, G. Cao,
889
+ and G. Guo, An automated approach for consecutive tun-
890
+ ing of quantum dot arrays, Applied Physics Letters 121,
891
+ 084002 (2022), https://doi.org/10.1063/5.0111128.
892
+ [27] J. Ziegler, T. McJunkin, E. S. Joseph, S. S. Kalantre,
893
+ B. Harpt, D. E. Savage, M. G. Lagally, M. A. Eriksson,
894
+ J. M. Taylor, and J. P. Zwolak, Toward robust autotuning
895
+ of noisy quantum dot devices, Phys. Rev. Applied 17,
896
+ 024069 (2022).
897
+ [28] J. Darulov´a, M. Troyer, and M. C. Cassidy, Evaluation
898
+ of synthetic and experimental training data in super-
899
+ vised machine learning applied to charge-state detection
900
+ of quantum dots, Mach. Learn.: Sci. Technol. 2, 045023
901
+ (2021).
902
+ [29] J.
903
+ Ziegler,
904
+ F.
905
+ Luthi,
906
+ M.
907
+ Ramsey,
908
+ F.
909
+ Borjans,
910
+ G. Zheng, and J. P. Zwolak, Tuning arrays with rays:
911
+ Physics-informed tuning of quantum dot charge states,
912
+ arXiv:2209.03837 10.48550/arxiv.2209.03837 (2022).
913
+ [30] T. Thorbeck and N. M. Zimmerman, Determining
914
+ the location and cause of unintentional quantum dots
915
+ in a nanowire, J. Appl. Phys. 111, 064309 (2012),
916
+ https://doi.org/10.1063/1.3692387.
917
+ [31] M. W. Keller, J. M. Martinis, N. M. Zimmerman, and
918
+ A. H. Steinbach, Accuracy of electron counting using a
919
+ 7-junction electron pump, Applied Physics Letters 69,
920
+ 1804 (1996), publisher: American Institute of Physics.
921
+ [32] T.-K. Hsiao, C. van Diepen, U. Mukhopadhyay, C. Re-
922
+ ichl, W. Wegscheider, and L. Vandersypen, Efficient or-
923
+ thogonal control of tunnel couplings in a quantum dot
924
+ array, Phys. Rev. Applied 13, 054018 (2020).
925
+ [33] H. Qiao, Y. P. Kandel, K. Deng, S. Fallahi, G. C. Gard-
926
+ ner, M. J. Manfra, E. Barnes, and J. M. Nichol, Coher-
927
+ ent multispin exchange coupling in a quantum-dot spin
928
+ chain, Phys. Rev. X 10, 031006 (2020).
929
+ [34] R. Hanson, L. P. Kouwenhoven, J. R. Petta, S. Tarucha,
930
+ and L. M. K. Vandersypen, Spins in few-electron quan-
931
+ tum dots, Rev. Mod. Phys. 79, 1217 (2007).
932
+ [35] T. Lin, P. Dollar, R. Girshick, K. He, B. Hariharan, and
933
+ S. Belongie, Feature pyramid networks for object detec-
934
+ tion, in 2017 IEEE Conference on Computer Vision and
935
+ Pattern Recognition (CVPR) (IEEE Computer Society,
936
+ Los Alamitos, CA, USA, 2017) pp. 936–944.
937
+ [36] P. Virtanen, R. Gommers, T. E. Oliphant, M. Haber-
938
+ land, T. Reddy, D. Cournapeau, E. Burovski, P. Pe-
939
+ terson, W. Weckesser, J. Bright, S. J. van der Walt,
940
+ M. Brett, J. Wilson, K. J. Millman, N. Mayorov, A. R. J.
941
+ Nelson, E. Jones, R. Kern, E. Larson, C. J. Carey, ˙I. Po-
942
+ lat, Y. Feng, E. W. Moore, J. VanderPlas, D. Laxalde,
943
+ J. Perktold, R. Cimrman, I. Henriksen, E. A. Quintero,
944
+ C. R. Harris, A. M. Archibald, A. H. Ribeiro, F. Pe-
945
+ dregosa, P. van Mulbregt, and SciPy 1.0 Contributors,
946
+ SciPy 1.0: Fundamental Algorithms for Scientific Com-
947
+ puting in Python, Nature Methods 17, 261 (2020).
948
+ [37] Student, The probable error of a mean, Biometrika 6, 1
949
+ (1908).
950
+ [38] Calculating the error of the pixel classifier directly is
951
+ made difficult by necessary but imperfect resizing of la-
952
+ bels on evaluation.
953
+ [39] O. Ronneberger, P. Fischer, and T. Brox, U-net: Convo-
954
+ lutional networks for biomedical image segmentation, in
955
+ Medical Image Computing and Computer-Assisted Inter-
956
+ vention – MICCAI 2015, edited by N. Navab, J. Horneg-
957
+ ger, W. M. Wells, and A. F. Frangi (Springer Interna-
958
+ tional Publishing, Cham, 2015) pp. 234–241.
959
+ [40] C. C. Clogg, E. Petkova, and A. Haritou, Statistical
960
+ methods for comparing regression coefficients between
961
+ models, American Journal of Sociology 100, 1261 (1995),
962
+ https://doi.org/10.1086/230638.
963
+ [41] J. P. Zwolak, S. S. Kalantre, T. McJunkin, B. J. We-
964
+ ber, and J. M. Taylor, Ray-based classification framework
965
+ for high-dimensional data, in Third Workshop on Ma-
966
+ chine Learning and the Physical Sciences (NeurIPS 2020)
967
+ (Vancouver, Canada, 2020) pp. 1–7, arXiv:2010.00500.
968
+ [42] A. Chatterjee, F. Ansaloni, T. Rasmussen, B. Brovang,
969
+ F. Fedele, H. Bohuslavskyi, O. Krause, and F. Kuem-
970
+ meth,
971
+ Autonomous
972
+ estimation
973
+ of
974
+ high-dimensional
975
+ coulomb diamonds from sparse measurements, Phys.
976
+ Rev. Appl. 18, 064040 (2022).
977
+
AtFAT4oBgHgl3EQfrx4U/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
INAyT4oBgHgl3EQfffiq/content/tmp_files/2301.00342v1.pdf.txt ADDED
@@ -0,0 +1,855 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Many-body collective neutrino oscillations:
2
+ recent developments
3
+ Amol V. Patwardhan, Michael J. Cervia, Ermal Rrapaj, Pooja Siwach,
4
+ A. B. Balantekin
5
+ Abstract Neutrino flavor transformations in core-collapse supernovae and binary
6
+ neutron star mergers represent a complex and unsolved problem that is integral to
7
+ our understanding of the dynamics and nucleosynthesis in these environments. The
8
+ high number densities of neutrinos present in these environments can engender var-
9
+ ious collective effects in neutrino flavor transformations, driven either by neutrino-
10
+ neutrino coherent scattering, or in some cases, through collisional (incoherent) in-
11
+ teractions. An ensemble of neutrinos undergoing coherent scattering among them-
12
+ selves is an interacting quantum many-body system—as such, there is a tantalising
13
+ prospect of quantum entanglement developing between the neutrinos, which can
14
+ leave imprints on their flavor evolution histories. Here, we seek to summarize re-
15
+ cent progress that has been made towards understanding this phenomenon.
16
+ Amol V. Patwardhan
17
+ SLAC National Accelerator Laboratory, 2575 Sand Hill Rd, Menlo Park, CA 94025
18
+ e-mail: [email protected]
19
+ Michael J. Cervia
20
+ George Washington University, 725 21st St NW, Washington, DC 20052
21
+ e-mail: [email protected]
22
+ Ermal Rrapaj
23
+ University of California, Berkeley, CA 94720-7300
24
+ e-mail: [email protected]
25
+ Pooja Siwach
26
+ University of Wisconsin, 1150 University Ave, Madison, WI 53706
27
+ e-mail: [email protected]
28
+ A.B. Balantekin
29
+ University of Wisconsin, 1150 University Ave, Madison, WI 53706
30
+ e-mail: [email protected]
31
+ 1
32
+ arXiv:2301.00342v1 [hep-ph] 1 Jan 2023
33
+
34
+ 2
35
+ Amol V. Patwardhan, Michael J. Cervia, Ermal Rrapaj, Pooja Siwach, A. B. Balantekin
36
+ Motivation: Supernovae, Mergers, and the Early Universe
37
+ In extreme astrophysical environments such as core-collapse supernova explosions,
38
+ and binary neutron star (or black hole - neutron star) mergers, as well as during
39
+ certain epochs in the early universe, neutrinos dominate the transport of energy,
40
+ entropy, and lepton number (for example, see [Janka et al., 2007, Burrows and
41
+ Vartanyan, 2021, Fuller and Haxton, 2022, Foucart, 2022, Kyutoku et al., 2018,
42
+ Grohs et al., 2016], etc.). The key processes governing neutrino transport in these
43
+ environments are electron neutrino and antineutrino captures on nucleons, i.e.,
44
+ νe +n ⇌ p+e−
45
+ (1)
46
+ ¯νe + p ⇌ n+e+
47
+ (2)
48
+ A consequence of the typical temperatures and densities of these environments is
49
+ that neutrinos decouple with energies of O(1–10)MeV, and therefore, the µ and τ
50
+ flavor (anti-)neutrinos are unable to participate in these charged-current processes,
51
+ due to there not being enough energy to produce µ and τ leptons in the final state.
52
+ Given the importance of these processes in the energy transport, as well as in de-
53
+ termining the neutron-to-proton ratio and the resulting nucleosynthesis prospects
54
+ (e.g., [Surman and McLaughlin, 2004, Mart´ınez-Pinedo et al., 2017, Kajino et al.,
55
+ 2014, Fr¨ohlich et al., 2015, Langanke et al., 2019, Roberts et al., 2017, Steigman,
56
+ 2012, Grohs et al., 2016]), the flavor-asymmetric nature of charged-current capture
57
+ necessitates a thorough understanding of neutrino flavor evolution in these envi-
58
+ ronments. The potential impact of neutrino flavor evolution on nucleosynthesis has
59
+ already been studied in various contexts (e.g., [Qian et al., 1993, Yoshida et al.,
60
+ 2006, Duan et al., 2011, Kajino et al., 2012, Wu et al., 2015, Sasaki et al., 2017, Bal-
61
+ antekin, 2018, Xiong et al., 2019, Xiong et al., 2020]).
62
+ In what follows, we shall summarize recent progress in our understanding of
63
+ a particular facet of neutrino oscillations in extreme astrophysical environments—
64
+ namely, the quantum many-body nature of collective neutrino oscillations engen-
65
+ dered by ν-ν interactions in dense neutrino streams.
66
+ Introduction to collective neutrino oscillations
67
+ The neutral current weak term of the Standard Model (SM) allows neutrinos to
68
+ interact pairwise via virtual Z-boson exchange or, more simply, in the low-energy
69
+ effective theory, via the Fermi four-point interaction
70
+ Hint ≡ GF
71
+
72
+ 2 ∑
73
+ f,g
74
+ νgγµνgν f γµνf ,
75
+ (3)
76
+ where f,g span the flavor state indices. The relevance of these interactions in en-
77
+ vironments where the number densities of neutrinos are comparable to (or larger
78
+
79
+ Many-body collective neutrino oscillations: recent developments
80
+ 3
81
+ than) those of charged leptons, e.g., in core-collapse supernovae, binary neutron
82
+ star mergers, as well as in the early universe, had been discussed in [Notzold and
83
+ Raffelt, 1988, Fuller et al., 1987]. But the extent of their importance in changing
84
+ the flavor content of neutrinos, via diagonal and off-diagonal contributions to the
85
+ neutrino Hamiltonian, was not fully recognized until later [Pantaleone, 1992a, Pan-
86
+ taleone, 1992b, Samuel, 1993].
87
+ Considering pairs of neutrinos with well-defined incoming momenta p and q
88
+ (i.e., plane wave states) and the same pair of outgoing momenta (i.e., “forward
89
+ scattering” neutrinos, the contributions of which can be added coherently), the off-
90
+ diagonal matrix elements of the interaction Hamiltonian Hint may be interpreted as
91
+ arising from “flavor swaps” between neutrino pairs (in the flavor basis). Because
92
+ the off-diagonal term exchanges flavor between the “test” and the “background”
93
+ neutrinos, the flavor evolution of the interacting neutrinos constitutes a many-body
94
+ problem, potentially rendering the one-particle propagation formalism [Samuel,
95
+ 1993, Sigl and Raffelt, 1993, Qian and Fuller, 1995] inadequate for describing
96
+ the resulting dynamics. Notably, the interaction Hamiltonian Hint does not com-
97
+ mute with the Hamiltonian terms corresponding to flavor oscillations in vacuum
98
+ and neutrino interactions with background matter. Consequently, in a regime where
99
+ the strength of these terms is comparable in scale to the ν-ν interaction strength,
100
+ diagonalizing this Hamiltonian is not straightforward and the many-body problem
101
+ acquires a nontrivial nature. Here, the entire Hilbert space of N interacting neutrinos
102
+ and antineutrinos in nf flavors has dimension nN
103
+ f .
104
+ Despite emphasis on the high nonlinearity of this problem, [Samuel, 1993] had
105
+ proposed that a statistical mechanical approach, whereby a two-flavor neutrino den-
106
+ sity matrix is treated as interacting with a background of neutrinos and antineutri-
107
+ nos, could describe the evolution of a dense neutrino gas for certain portions of this
108
+ parameter space. This analysis was extended by [Sigl and Raffelt, 1993] to nf ≥ 2
109
+ flavors with proposed evolution of nf ×n f density matrices via quantum Boltzmann
110
+ equations, including collision integrals as well as more general, potentially non-SM
111
+ coupling between flavors. In these treatments, the collisional contributions can lead
112
+ to a nontrivial loss of coherence being reflected in the density matrices of individual
113
+ neutrinos. However, the ability to calculate multi-body wave functions that exhibit
114
+ ν-ν correlations is relinquished, in exchange for a more favourable scaling of com-
115
+ putational complexity with the number of neutrinos in the simulation. Along these
116
+ lines, [Qian and Fuller, 1995] proposed a physical ansatz that the wave function of
117
+ the ensemble is not a coherent many-body state, but simply composed of single-
118
+ neutrino wave functions with random relative phases, to be summed incoherently,
119
+ called the Random Phase Approximation (RPA). In this way, each neutrino density
120
+ matrix is taken to be pure, and the effective Hilbert space dimension is reduced to
121
+ nf N. In kind, the complexity of collective oscillations calculations becomes greatly
122
+ simplified. This ansatz amounts to a “mean field approximation” wherein expecta-
123
+ tion values of operator products may be replaced by products of the individual op-
124
+ erator expectation values. Notably, this physical description of neutrinos expressly
125
+ prohibits the quantum entanglement between neutrinos. As such, assessing the va-
126
+ lidity of this ansatz involves determining the extent to which quantum effects are
127
+
128
+ 4
129
+ Amol V. Patwardhan, Michael J. Cervia, Ermal Rrapaj, Pooja Siwach, A. B. Balantekin
130
+ needed to correct this approximation. In this chapter, we discuss recent progress
131
+ along this front.
132
+ Before delving into the chapter, we mention in passing that recent years have seen
133
+ a rapid growth of interest in flavor instabilities and resulting fast flavor oscillation,
134
+ even within the scope of the mean field approximation. For more information we
135
+ refer the reader to the chapter on “Fast Flavor Transformations” by [Richers and Sen,
136
+ 2022], or the review articles by [Chakraborty et al., 2016, Tamborra and Shalgar,
137
+ 2021].
138
+ Many-body Hamiltonian for interacting neutrinos
139
+ The Hamiltonian describing a system of interacting neutrinos can be written in terms
140
+ of generators of SU(nf ), and it possesses a SU(nf ) rotation symmetry in neutrino
141
+ flavor. A significant feature of ν-ν interactions is the dependence of the interac-
142
+ tion strength on the intersection angle between their trajectories. This dependence
143
+ introduces a geometric complexity to the problem, in addition to the complexity
144
+ associated with the exponential scaling of the Hilbert space.
145
+ For simplicity, if we consider neutrino mixing between only two flavors, νe and
146
+ νx, then a Hamiltonian consisting of terms that represent vacuum mixing as well as
147
+ ν-ν interactions can be written as
148
+ H = ∑
149
+ p
150
+ ωp⃗B· ⃗Jp +
151
+
152
+ 2GF
153
+ V
154
+
155
+ p,q
156
+ (1−�p·�q) ⃗Jp · ⃗Jq ,
157
+ (4)
158
+ where ⃗B = (0,0,−1) in the mass-basis representation, and ωp = δm2/(2|p|) are
159
+ the vacuum oscillation frequencies for neutrinos with momenta p. Here, �p and �q
160
+ are the unit vectors along the momenta of the interacting neutrinos, and V is the
161
+ quantization volume. For ease of notation, one can define a ν-ν coupling parameter
162
+ µ ≡
163
+
164
+ 2GFN/V, where N is the total number of interacting neutrinos. The oper-
165
+ ators ⃗Jp represent the neutrino “isospin” in the mass basis, where isospin up and
166
+ down correspond to the mass basis states |ν1⟩ and |ν2⟩. In this depiction, ⃗B can be
167
+ interpreted as a “background field” with which the neutrino isospins interact. Here,
168
+ we exclude the term representing neutrino interactions with ordinary matter (e.g.,
169
+ charged leptons), since it has a structure that is conceptually similar to the vacuum
170
+ oscillation term—i.e., consisting of individual neutrinos interacting with a back-
171
+ ground. In contrast, the ν-ν interaction term consists of pairs of neutrino isospins
172
+ interacting with each other.
173
+ In terms of the Fermionic creation and annihilation operators, the neutrino
174
+ isospins are described as [Balantekin and Pehlivan, 2006]
175
+ J+
176
+ p = a†
177
+ 1(p)a2(p) ,
178
+ Jz
179
+ p = 1
180
+ 2
181
+
182
+ a†
183
+ 1 (p)a1(p)−a†
184
+ 2 (p)a2(p)
185
+
186
+ ,
187
+ (5)
188
+
189
+ Many-body collective neutrino oscillations: recent developments
190
+ 5
191
+ with J−
192
+ p = (J+
193
+ p )†. In the spin-1/2 representation, one can write the isospin operators
194
+ in terms of Pauli matrices: i.e., ⃗Jp = ⃗σp/2, where σp is a vector of Pauli matrices
195
+ defined in the subspace of the neutrino with momentum p.
196
+ Path integral formulation
197
+ An assessment of quantum corrections to a mean field picture can in principle be
198
+ performed via a coherent state analysis, as formulated by [Balantekin and Pehlivan,
199
+ 2006]. Schematically, in this procedure, one seeks to calculate the matrix elements
200
+ of the time evolution operator U(tf ;ti) for a single neutrino in the basis of SU(nf )
201
+ coherent states |z⟩ for neutrinos (and/or antineutrinos) in n f flavors, equivalent to a
202
+ path integral
203
+ ⟨zf |U(t f ;ti)|zi⟩ =
204
+
205
+ D[z,z∗] exp(iS[z,z∗])
206
+ (6)
207
+ of the derived action
208
+ S[z,z∗] =
209
+ � tf
210
+ ti
211
+ dt
212
+
213
+ ⟨z(t)|(i∂t −H)|z(t)⟩−ilog⟨z f |zi⟩
214
+
215
+ ,
216
+ (7)
217
+ where H is the Hamiltonian of the many-body system. A saddle-point approxima-
218
+ tion of the resulting path integral yields the classical action, which is in complete
219
+ agreement with the RPA used to derive the mean field theory for collective neutrino
220
+ oscillations. However, in this perspective, analyzing quantum corrections to this ap-
221
+ proximation entails a careful analysis of the Hessian matrix of the action integral
222
+ derived from this procedure. Such mathematical analysis has not yet been presented
223
+ to date.
224
+ Beyond the Mean-Field: Entanglement, Correlations, and
225
+ Dynamical Phase Transitions
226
+ Early literature
227
+ The seminal work describing the ν-ν interaction Hamiltonian from Eq. (3) recog-
228
+ nized that these interactions give rise to a quantum many-body problem, which may
229
+ not in the general case be factorizable in terms of a one-particle effective approxi-
230
+ mation [Pantaleone, 1992a, Pantaleone, 1992b]. Subsequently, there were some at-
231
+ tempts to ascertain the validity of the one-particle effective approximation [Bell
232
+ et al., 2003, Friedland and Lunardini, 2003b, Friedland and Lunardini, 2003a, Fried-
233
+ land et al., 2006]. In these works, the flavor evolution of interacting neutrinos was
234
+ analyzed with two different approaches: (i) using two intersecting beams of neutri-
235
+
236
+ 6
237
+ Amol V. Patwardhan, Michael J. Cervia, Ermal Rrapaj, Pooja Siwach, A. B. Balantekin
238
+ nos, where the flavor evolution was described in terms of a sequence of elementary
239
+ scattering amplitudes, and (ii) using a neutrino ensemble represented as interacting
240
+ plane waves in a box.
241
+ Following initial disagreement regarding whether substantial quantum entangle-
242
+ ment can develop among interacting neutrinos [Bell et al., 2003, Friedland and Lu-
243
+ nardini, 2003b], it was subsequently concluded that the build-up of entanglement
244
+ and resulting flavor conversion would occur on timescales whose scaling is sugges-
245
+ tive of incoherent effects [Friedland and Lunardini, 2003a]. These conclusions were
246
+ further generalized in [Friedland et al., 2006]. However, these analyses nevertheless
247
+ involved several simplifications, most notably, the omission of the one-body terms
248
+ in the Hamiltonian. The interplay between vacuum oscillations and ν-ν interaction
249
+ terms has been shown to give rise to interesting collective phenomena such as “spec-
250
+ tral splits” [Duan et al., 2006a, Duan et al., 2006b, Duan et al., 2007b, Raffelt and
251
+ Smirnov, 2007b, Raffelt and Smirnov, 2007a], even in the mean-field approxima-
252
+ tion. Therefore, studying the quantum many-body dynamics of collective neutrino
253
+ oscillations, with both one- and two-body terms fully incorporated, remains an in-
254
+ teresting problem.
255
+ With these seemingly conflicting results in the past predicting either a vanish-
256
+ ingly small contribution in the large system size limit [Friedland and Lunardini,
257
+ 2003a, Friedland et al., 2006] or substantial flavor evolution over time scales τF ∼
258
+ µ−1 log(N) that can remain relevant for large systems [Bell et al., 2003, Sawyer,
259
+ 2004], the role of entanglement and quantum effects in the out-of-equilibrium dy-
260
+ namics [Eisert et al., 2015] of neutrinos has received renewed interest recently (e.g.,
261
+ [Cervia et al., 2019, Rrapaj, 2020] and subsequent works mentioned later in this
262
+ chapter). Note that flavor oscillations on the time scale τF can be considered to be
263
+ “fast”, different from “slow” oscillations occurring over τL ∼ µ−1√
264
+ N. In the lit-
265
+ erature on collective flavor effects in the mean field approximation, one can more
266
+ commonly find “fast” and “slow” oscillations associated with time scales ∼ µ−1
267
+ and ∼ √µω (or ω), respectively.
268
+ Single-angle approximation, invariants, and integrability
269
+ To circumvent the geometric complexity of the problem, the frequently-employed
270
+ single-angle approximation replaces the angle-dependent (i.e., �p,�q-dependent) in-
271
+ teraction strengths among pairs of neutrinos with a single, appropriately chosen
272
+ classical average over the various neutrino trajectories. In this limit, one can de-
273
+ fine a trajectory-averaged interaction parameter µ ≡ (
274
+
275
+ 2GFN/V)⟨1 − �p · �q⟩, and
276
+ approximate the Hamiltonian as
277
+ H = ∑
278
+ ωp
279
+ ωp⃗B· ⃗Jωp + µ
280
+ N
281
+ ⃗J · ⃗J ,
282
+ (8)
283
+
284
+ Many-body collective neutrino oscillations: recent developments
285
+ 7
286
+ where ⃗J = ∑ωp ⃗Jωp is the total neutrino isospin. Note that, in this limit, the neutrino
287
+ flavor state becomes trajectory-independent, introducing a considerable simplifica-
288
+ tion in the problem. As a result, the neutrinos may be indexed simply by the mag-
289
+ nitudes of their momenta (or equivalently, by their vacuum oscillation frequencies
290
+ ωp), rather than by the momenta themselves (magnitude and direction). The ν-ν
291
+ coupling in general will depend on time. In the context of supernovae, a commonly
292
+ employed expression for µ is derived from the spherically symmetric single-angle
293
+ neutrino bulb model, first described in [Duan et al., 2006c]:
294
+ µ(r) = µ0
295
+
296
+ �1−
297
+
298
+ 1−
299
+ �Rν
300
+ r
301
+ �2
302
+
303
+
304
+ 2
305
+ ,
306
+ (9)
307
+ where r is the distance from the center of a “neutrino-sphere” of radius Rν, which
308
+ represents a sharp surface where neutrinos decouple from nuclear matter and be-
309
+ gin free streaming outwards from the proto-neutron star. We also define µ0 ≡
310
+ (GF/
311
+
312
+ 2)(N/V) = µ(Rν) to be the interaction strength at the neutrino-sphere. Here,
313
+ the neutrino emission is assumed to be time-invariant over the short time scales as-
314
+ sociated with neutrino propagation through the supernova envelope, so the interac-
315
+ tion strength depends explicitly only on position, rather than time. In the neutrino-
316
+ driven wind phase of core-collapse supernovae, which occurs over a time window
317
+ of O(1–10) s after core bounce, one may expect Rν ≃ 20km and µ0 ∼ 105ω0, where
318
+ ω0 ∼ 10−16 MeV is the scale of the vacuum oscillations. During the shock breakout
319
+ or “neutronization burst” phase that occurs earlier, around 10 ms after core bounce,
320
+ the proto-neutron star can be more extended, with Rν ≳ 50–60 km, but the neutrino
321
+ luminosity is also much higher, resulting in µ0 ∼ 106ω0.
322
+ It has been shown that a single-angle Hamiltonian describing neutrino mixing
323
+ in vacuum and ν-ν interactions possesses a number of conserved charges which
324
+ commute with the Hamiltonian [Pehlivan et al., 2011]. These are analogous to
325
+ the “Gaudin magnets” [Gaudin, M., 1976] that had been previously identified as
326
+ the conserved charges of the pairing-force Hamiltonian in nuclear and condensed-
327
+ matter physics [Richardson, 1963, Richardson and Sherman, 1964, Richardson,
328
+ 1965]. These conserved charges are related to the integrability of the Hamiltonian—
329
+ meaning that it is possible to obtain, in principle, exact eigenvalues and eigenstates
330
+ of this Hamiltonian in terms of closed-form solutions to a set of algebraic “Bethe-
331
+ Ansatz” equations [Bethe, 1931]. Based on these ideas, specific procedures for the
332
+ eigen-decomposition of a single-angle neutrino Hamiltonian have been outlined in
333
+ the literature [Pehlivan et al., 2011, Birol et al., 2018, Patwardhan et al., 2019].
334
+ Besides descriptions in terms of instantaneously conserved charges, analogies
335
+ with other many-body problems have been fruitful to yield an explanation of the
336
+ neutrino flavor spectral split in terms of a Bardeen-Cooper-Schrieffer (BCS)-Bose-
337
+ Einstein Condensate (BEC) crossover-like phenomenon [Pehlivan et al., 2017], as
338
+ well as to help provide many-body predictions of a spectral split [Birol et al., 2018]
339
+ specifically in the case of an initial many-body wave function with all neutrinos in
340
+ the electron flavor state.
341
+
342
+ 8
343
+ Amol V. Patwardhan, Michael J. Cervia, Ermal Rrapaj, Pooja Siwach, A. B. Balantekin
344
+ Instabilities and dynamical phase transitions
345
+ Collective neutrino oscillations are generally assumed to be caused by unstable
346
+ modes in the mean field dynamics generated by the Hamiltonian described in
347
+ Eq. (4) (for two flavors). These instabilities are able to amplify initially small
348
+ flavor perturbations exponentially fast (e.g., [Sawyer, 2004, Sawyer, 2005, Duan
349
+ et al., 2010, Chakraborty et al., 2016, Izaguirre et al., 2017, Tamborra and Shalgar,
350
+ 2021, Richers and Sen, 2022] and references therein). The presence of the forward-
351
+ scattering interaction can allow collective effects to develop when µ ≳ ωp, giving
352
+ rise to interesting phenomena like synchronization [Pastor et al., 2002, Fuller and
353
+ Qian, 2006, Raffelt and Tamborra, 2010, Akhmedov and Mirizzi, 2016], bipolar os-
354
+ cillations [Kosteleck´y and Samuel, 1995, Duan et al., 2006c, Duan et al., 2007a] and
355
+ spectral splits/swaps [Duan et al., 2006b, Duan et al., 2007b, Raffelt and Smirnov,
356
+ 2007b, Dasgupta et al., 2009, Martin et al., 2020].
357
+ On the other hand, in descriptions of interacting neutrino systems that per-
358
+ mit many-body quantum dynamics, oscillations that develop on “fast” timescales
359
+ are generally associated with rapid dynamical development of the neutrino en-
360
+ tanglement entropy [Cervia et al., 2019, Rrapaj, 2020, Roggero, 2021a, Roggero,
361
+ 2021b, Patwardhan et al., 2021]. The dynamically generated entanglement between
362
+ neutrinos is seen to be correlated with deviations from the mean-field dynamics of
363
+ the system [Cervia et al., 2019, Rrapaj, 2020] and with the presence of spectral splits
364
+ in the neutrino energy distributions [Patwardhan et al., 2021]. An example of such
365
+ a calculation is depicted in Fig. 1. In [Roggero et al., 2022], rapid entanglement and
366
+ mean field instabilities were also found to be linked for certain angular setups.
367
+ As shown in [Roggero, 2021a, Roggero, 2021b] in the single angle approxima-
368
+ tion, when the frequency difference between two neutrino beams (δω) is positive
369
+ and comparable to the ν-ν interaction coupling (µ), 0 < δω ≲ µ, rapid and strong
370
+ flavor oscillations develop. This rather particular finding can be understood in terms
371
+ of the presence of a Dynamic Phase Transition (DPT) [Heyl et al., 2013, Heyl,
372
+ 2018], which can be characterized by the introduction of the Loschmidt echo,
373
+ L (t) = |⟨Φ|exp(−itH)|Φ⟩|2 ,
374
+ (10)
375
+ with |Φ⟩ the initial state at t = 0. The quantity L (t) is a fidelity measure [Gorin
376
+ et al., 2006] that quantifies the probability for the system to return to its initial state.
377
+ A DPT is then characterized by non-analyticities in the rate function
378
+ λ(t) = − 1
379
+ N log[L (t)] ,
380
+ (11)
381
+ where N is the total number of particles in the system and λ(t) an intensive “free
382
+ energy” [Heyl et al., 2013, Gambassi and Silva, 2012]. Here, the rate λ(t) plays the
383
+ role of a non-equilibrium equivalent of the thermodynamic free-energy. Notably,
384
+ other definitions of DPT are possible, for instance, time-averaged order parame-
385
+ ters [Sciolla and Biroli, 2011, Sciolla and Biroli, 2013, ˇZunkoviˇc et al., 2018].
386
+
387
+ Many-body collective neutrino oscillations: recent developments
388
+ 9
389
+ −1
390
+ −0.8
391
+ −0.6
392
+ −0.4
393
+ −0.2
394
+ 0
395
+ 0.2
396
+ 0.4
397
+ 0.6
398
+ 0.8
399
+ 1
400
+ 200
401
+ 500
402
+ 1000
403
+ 2000
404
+ P MB
405
+ z
406
+ (ωp)
407
+ r (in units of ω−1
408
+ 0 )
409
+ Pz(ω1)
410
+ Pz(ω2)
411
+ Pz(ω3)
412
+ Pz(ω4)
413
+ Pz(ω5)
414
+ Pz(ω6)
415
+ Pz(ω7)
416
+ Pz(ω8)
417
+ −1
418
+ −0.8
419
+ −0.6
420
+ −0.4
421
+ −0.2
422
+ 0
423
+ 0.2
424
+ 0.4
425
+ 0.6
426
+ 0.8
427
+ 1
428
+ 200
429
+ 500
430
+ 1000
431
+ 2000
432
+ P MF
433
+ z
434
+ (ωp)
435
+ r (in units of ω−1
436
+ 0 )
437
+ Pz(ω1)
438
+ Pz(ω2)
439
+ Pz(ω3)
440
+ Pz(ω4)
441
+ Pz(ω5)
442
+ Pz(ω6)
443
+ Pz(ω7)
444
+ Pz(ω8)
445
+ 0
446
+ 0.2
447
+ 0.4
448
+ 0.6
449
+ 0.8
450
+ 1
451
+ 200
452
+ 500
453
+ 1000
454
+ 2000
455
+ S(ωp)
456
+ r (in units of ω−1
457
+ 0 )
458
+ S(ω1)
459
+ S(ω2)
460
+ S(ω3)
461
+ S(ω4)
462
+ S(ω5)
463
+ S(ω6)
464
+ S(ω7)
465
+ S(ω8)
466
+ −1
467
+ −0.5
468
+ 0
469
+ 0.5
470
+ 1
471
+ 1
472
+ 2
473
+ 3
474
+ 4
475
+ 5
476
+ 6
477
+ 7
478
+ 8
479
+ 0
480
+ 0.2
481
+ 0.4
482
+ 0.6
483
+ Pz(ωp)
484
+ S(ωp)
485
+ ω (in units of ω0)
486
+ Pz (initial)
487
+ P MB
488
+ z
489
+ (nal)
490
+ P MF
491
+ z
492
+ (nal)
493
+ S(ωp) (nal)
494
+ Fig. 1 Evolution of an initial state |νe⟩⊗4 |νx⟩⊗4 from a starting radius r0 such that µ(r0) = 5ω0,
495
+ with a small mixing angle (θ = 0.161) and discrete, equally spaced oscillation frequencies
496
+ ωk = kω0, and a time-varying neutrino interaction strength µ(r) motivated by the neutrino bulb
497
+ model [Duan et al., 2006b], in the single-angle approximation according to Eqs. (8) and (9). Details
498
+ of this calculation can be found in [Cervia et al., 2019]. Top left: Evolution of the z-components
499
+ of the neutrino isospin expectation values (also known as “Polarization vectors”) in the mass basis,
500
+ i.e., Pz ≡ 2⟨Jz⟩, for the full many-body quantum system. Top right: Same as top left, but in the
501
+ mean-field approximation. Bottom left: Evolution of the entanglement entropy of each neutrino,
502
+ with respect to the rest of the ensemble. Bottom right: Asymptotic values of Pz vs ωk, in the full
503
+ many-body calculation (purple), and in the mean-field approximation (green), together with the
504
+ initial Pz values (red), and the asymptotic entanglement entropies (dark orange). Neutrinos located
505
+ closest to the spectral splits in the energy distributions (in this case, at ω2 and ω7) develop the
506
+ largest amount of entanglement and thereby experience the most significant deviations compared
507
+ to their mean-field evolution.
508
+ Phase-space analysis
509
+ In a recent work [Lacroix et al., 2022], this problem was further explored by ana-
510
+ lyzing the evolution of neutrino flavor and entanglement in phase space. The setup
511
+ consisted of two sets (beams) of neutrinos interacting with each other. In this anal-
512
+ ysis, the Husimi quasi-probability or “Q” representation [Husimi, 1940] was con-
513
+ structed for the reduced density operator of neutrinos in one of the beams, using an
514
+ over-complete basis of coherent states. In the limit of infinite neutrino number, the
515
+ Q representation acquires the interpretation of a classical phase-space probability
516
+ distribution.
517
+
518
+ 10
519
+ Amol V. Patwardhan, Michael J. Cervia, Ermal Rrapaj, Pooja Siwach, A. B. Balantekin
520
+ For this two-beam interacting neutrino system, it was demonstrated that, while
521
+ at early times the quasi-probability distribution remains relatively localized, at late
522
+ times it develops a multi-modal structure with several localized peaks. This delocal-
523
+ ization is indicative of non-Gaussian entanglement, which suggests that any approx-
524
+ imate method beyond the mean-field relying on only the first and second moments
525
+ of neutrino observables may not be sufficient to describe the long-term evolution
526
+ of this system. Based on the phase space analysis, a new method for approximat-
527
+ ing the exact evolution of the interacting neutrino system was proposed, wherein
528
+ the quantum mechanical many-body evolution is replaced by a statistical average of
529
+ ‘mean-field’ solutions, with a Gaussian distribution of initial conditions around the
530
+ exact starting point of the system [Lacroix and Ayik, 2014].
531
+ Compact Representations for studying many body effects
532
+ Still allowing for possibilities of mixed one-neutrino density matrices, one pro-
533
+ posal [Volpe et al., 2013] to determine quantum corrections is to systematically
534
+ incorporate n-body density matrices ρ1...n for n ≥ 1, given by
535
+ ρ1...n =
536
+ N!
537
+ (N −n)!Trn+1...Nρ1...N,
538
+ (12)
539
+ into the coupled equations of motion for N neutrinos, as follows:
540
+ i∂tρ1...n = [H1...n,ρ1...n]+
541
+ n
542
+
543
+ s=1
544
+ Trn+1[V(s,n+1),ρ1...n+1],
545
+ (13)
546
+ where H1...n is the Hamiltonian truncated for the first n neutrinos in a given ordering
547
+ and V(i, j) is the two-body interaction potential for a pair of neutrinos (i, j). This
548
+ procedure is based on the Bogoliubov-Born-Green-Kirkwood-Yvon (BBGKY) hi-
549
+ erarchy for density matrices. Here, the mean field theory interaction of neutrinos
550
+ and antineutrinos with the background gas is reproduced by restricting to n = 2
551
+ and estimating ρ12 ≈ ρ1ρ2 (i.e., requiring the two-body correlation function to be
552
+ zero) in this picture, in a sense as a loop Feynman diagram for neutrino propagation.
553
+ In principle, investigating the importance of quantum corrections would practically
554
+ entail checking for convergence of results for physical observables as the n-body
555
+ correlation functions are incorporated for progressively increasing values of n in the
556
+ BBGKY hierarchy.
557
+ Owing to the exponential growth in the Hilbert space, classical (conventional)
558
+ computers are unable to exactly simulate systems of more than ≃ 20 neutrinos.
559
+ To overcome this difficulty, one can resort to compact representations of the wave-
560
+ function through tensor network methods [Roggero, 2021a, Roggero, 2021b, Cervia
561
+ et al., 2022], and more specifically matrix product states [Vidal, 2003, Schollw¨ock,
562
+ 2011, Paeckel et al., 2019]. In simplified setups, these methods allow for the com-
563
+ putation of systems of hundreds of neutrinos. Alternatively, when considering very
564
+
565
+ Many-body collective neutrino oscillations: recent developments
566
+ 11
567
+ dense neutrino gases (vacuum oscillations can be ignored), methods based on gen-
568
+ eralized angular momentum representations, by analogy between two flavor oscilla-
569
+ tions and spin systems, can reach up to thousands of neutrinos and predict the ther-
570
+ modynamic limit [Friedland and Lunardini, 2003a, Friedland et al., 2006, Xiong,
571
+ 2022, Roggero et al., 2022].
572
+ In the case of time-dependent interaction strength and all-to-all ν-ν interactions,
573
+ the more sophisticated tensor network method, namely, the time-dependent varia-
574
+ tional principle (TDVP) method has been utilized in [Cervia et al., 2022]. These
575
+ techniques provided considerable computational benefit for an initial state with all
576
+ neutrinos in the same flavor, allowing for evolution of a system with ≈ 50 oscil-
577
+ lation modes. This was a consequence of the entanglement among neutrinos being
578
+ more localized in certain regions of the neutrino energy distribution. For systems
579
+ with initial states being a mixture of νe and νx flavors, the entanglement is more de-
580
+ localized, and therefore, the comparative advantage gained through TDVP methods
581
+ is less dramatic, although work remains in progress on this front.
582
+ For a general setup, quantum computers are a promising tool to solve the quan-
583
+ tum many-body problem. Initial steps [Hall et al., 2021, Yeter-Aydeniz et al.,
584
+ 2022, Illa and Savage, 2022, Amitrano et al., 2022] to simulate the collective neu-
585
+ trino oscillations on a quantum computer are already taken in this direction. In [Hall
586
+ et al., 2021] a sytem of four neutrinos was simulated on IBM’s quantum devices
587
+ using the real-time evolution. The unitary evolution operator U(t) = exp(−iHt)
588
+ was decomposed using the first order Trotter-Suzuki decomposition, where error
589
+ scales as O(t2). Since the interaction is long-range, a device with all-to-all con-
590
+ nectivity among qubits is preferred. As an alternative, SWAP operations have been
591
+ used to implement this interaction on a quantum device having connectivity among
592
+ neighboring qubits [Hall et al., 2021]. In [Yeter-Aydeniz et al., 2022], the hybrid
593
+ quantum-classical algorithm QLanczos (quantum Lanczos) was used to calculate
594
+ the eigenvalues of neutrino many-body interaction Hamiltonian [Patwardhan et al.,
595
+ 2019] on a quantum computer. Furthermore, the transition probabilities of collec-
596
+ tive neutrino oscillations were obtained by performing the real-time evolution using
597
+ trotterization. However, all these earlier quantum computing studies were limited to
598
+ a small system of four neutrinos due to constraints in the form of currently avail-
599
+ able quantum devices, which can perform only a limited number of operations with
600
+ low accuracy. More recently in [Amitrano et al., 2022], a trapped-ion quantum de-
601
+ vice was utilized to perform the simulations for up to eight neutrinos, thanks to the
602
+ all-to-all qubit connectivity in trapped-ion based architecture.
603
+ Concluding remarks
604
+ Studying the many-body quantum dynamics of dense neutrino systems remains an
605
+ active area of research, with various groups attempting to investigate the problem
606
+ using different types of classical and quantum computational tools, as well as ana-
607
+ lytic or semi-analytic descriptions. In environments where neutrinos are present in
608
+
609
+ 12
610
+ Amol V. Patwardhan, Michael J. Cervia, Ermal Rrapaj, Pooja Siwach, A. B. Balantekin
611
+ high number densities, they almost inevitably become the main carriers of energy
612
+ and lepton number, and as a result, the physics of neutrino flavor transformation
613
+ in these environments becomes particularly relevant for the dynamics and nucle-
614
+ osynthesis. Moreover, the close parallels between this problem and other quantum
615
+ many-body systems in nuclear and condensed-matter physics suggests that the re-
616
+ sults and insights obtained through these studies could have a much broader scope,
617
+ beyond just the field of neutrino physics.
618
+ References
619
+ Akhmedov and Mirizzi, 2016. Akhmedov, E. and Mirizzi, A. (2016). Another look at synchro-
620
+ nized neutrino oscillations. Nuclear Physics B, 908:382–407. Neutrino Oscillations: Celebrating
621
+ the Nobel Prize in Physics 2015.
622
+ Amitrano et al., 2022. Amitrano, V., Roggero, A., Luchi, P., Turro, F., Vespucci, L., and Pederiva,
623
+ F. (2022). Trapped-Ion Quantum Simulation of Collective Neutrino Oscillations. 2207.03189.
624
+ Balantekin, 2018. Balantekin, A. B. (2018). Collective Neutrino Oscillations and Nucleosynthe-
625
+ sis. AIP Conf. Proc., 1947(1):020012, 1710.04108.
626
+ Balantekin and Pehlivan, 2006. Balantekin, A. B. and Pehlivan, Y. (2006).
627
+ Neutrino–neutrino
628
+ interactions and flavour mixing in dense matter. Journal of Physics G: Nuclear and Particle
629
+ Physics, 34(1):47–65.
630
+ Bell et al., 2003. Bell, N. F., Rawlinson, A. A., and Sawyer, R. F. (2003). Speedup through entan-
631
+ glement: Many body effects in neutrino processes. Phys. Lett. B, 573:86–93, hep-ph/0304082.
632
+ Bethe, 1931. Bethe, H. (1931). Zur Theorie der Metalle. Zeitschrift fur Physik, 71(3-4):205–226.
633
+ Birol et al., 2018. Birol, S., Pehlivan, Y., Balantekin, A. B., and Kajino, T. (2018). Neutrino Spec-
634
+ tral Split in the Exact Many Body Formalism. Phys. Rev. D, 98(8):083002, 1805.11767.
635
+ Burrows and Vartanyan, 2021. Burrows, A. and Vartanyan, D. (2021). Core-Collapse Supernova
636
+ Explosion Theory. Nature, 589(7840):29–39, 2009.14157.
637
+ Cervia et al., 2019. Cervia, M. J., Patwardhan, A. V., Balantekin, A. B., Coppersmith, S. N., and
638
+ Johnson, C. W. (2019). Entanglement and collective flavor oscillations in a dense neutrino gas.
639
+ Phys. Rev. D, 100:083001.
640
+ Cervia et al., 2022. Cervia, M. J., Siwach, P., Patwardhan, A. V., Balantekin, A. B., Coppersmith,
641
+ S. N., and Johnson, C. W. (2022). Collective neutrino oscillations with tensor networks using a
642
+ time-dependent variational principle. Phys. Rev. D, 105:123025.
643
+ Chakraborty et al., 2016. Chakraborty, S., Hansen, R., Izaguirre, I., and Raffelt, G. (2016). Collec-
644
+ tive neutrino flavor conversion: Recent developments. Nucl. Phys. B, 908:366–381, 1602.02766.
645
+ Dasgupta et al., 2009. Dasgupta, B., Dighe, A., Raffelt, G. G., and Smirnov, A. Y. (2009). Multi-
646
+ ple spectral splits of supernova neutrinos. Phys. Rev. Lett., 103:051105.
647
+ Duan et al., 2011. Duan, H., Friedland, A., McLaughlin, G., and Surman, R. (2011). The influence
648
+ of collective neutrino oscillations on a supernova r-process. J. Phys. G, 38:035201, 1012.0532.
649
+ Duan et al., 2006a. Duan, H., Fuller, G. M., Carlson, J., and Qian, Y.-Z. (2006a). Coherent Devel-
650
+ opment of Neutrino Flavor in the Supernova Environment. Phys. Rev. Lett., 97:241101, astro-
651
+ ph/0608050.
652
+ Duan et al., 2006b. Duan, H., Fuller, G. M., Carlson, J., and Qian, Y.-Z. (2006b).
653
+ Simulation
654
+ of coherent nonlinear neutrino flavor transformation in the supernova environment: Correlated
655
+ neutrino trajectories. Phys. Rev. D, 74:105014.
656
+ Duan et al., 2007a. Duan, H., Fuller, G. M., Carlson, J., and Qian, Y.-Z. (2007a). Analysis of
657
+ collective neutrino flavor transformation in supernovae. Phys. Rev. D, 75:125005.
658
+ Duan et al., 2007b. Duan, H., Fuller, G. M., Carlson, J., and Qian, Y.-Z. (2007b).
659
+ Neutrino
660
+ mass hierarchy and stepwise spectral swapping of supernova neutrino flavors. Phys. Rev. Lett.,
661
+ 99:241802.
662
+
663
+ Many-body collective neutrino oscillations: recent developments
664
+ 13
665
+ Duan et al., 2006c. Duan, H., Fuller, G. M., and Qian, Y.-Z. (2006c). Collective neutrino flavor
666
+ transformation in supernovae. Phys. Rev. D, 74:123004.
667
+ Duan et al., 2010. Duan, H., Fuller, G. M., and Qian, Y.-Z. (2010).
668
+ Collective neu-
669
+ trino oscillations.
670
+ Annual Review of Nuclear and Particle Science, 60(1):569–594,
671
+ https://doi.org/10.1146/annurev.nucl.012809.104524.
672
+ Eisert et al., 2015. Eisert, J., Friesdorf, M., and Gogolin, C. (2015). Quantum many-body systems
673
+ out of equilibrium. Nature Physics, 11(2):124–130.
674
+ Foucart, 2022. Foucart, F. (2022). Neutrino transport in general relativistic neutron star merger
675
+ simulations. 2209.02538.
676
+ Friedland and Lunardini, 2003a. Friedland, A. and Lunardini, C. (2003a). Do many particle neu-
677
+ trino interactions cause a novel coherent effect? JHEP, 10:043, hep-ph/0307140.
678
+ Friedland and Lunardini, 2003b. Friedland, A. and Lunardini, C. (2003b). Neutrino flavor con-
679
+ version in a neutrino background: Single particle versus multiparticle description. Phys. Rev. D,
680
+ 68:013007, hep-ph/0304055.
681
+ Friedland et al., 2006. Friedland, A., McKellar, B. H. J., and Okuniewicz, I. (2006). Construc-
682
+ tion and analysis of a simplified many-body neutrino model. Phys. Rev. D, 73:093002, hep-
683
+ ph/0602016.
684
+ Fr¨ohlich et al., 2015. Fr¨ohlich, C., Casanova, J., Hempel, M., Liebend¨orfer, M., Melton, C. A.,
685
+ and Perego, A. (2015). Neutrinos and nucleosynthesis in core-collapse supernovae. AIP Conf.
686
+ Proc., 1604(1):178–184.
687
+ Fuller and Haxton, 2022. Fuller, G. M. and Haxton, W. C. (2022). Neutrinos in Stellar Astro-
688
+ physics. 2208.08050.
689
+ Fuller et al., 1987. Fuller, G. M., Mayle, R. W., Wilson, J. R., and Schramm, D. N. (1987). Reso-
690
+ nant Neutrino Oscillations and Stellar Collapse. Astrophys. J., 322:795.
691
+ Fuller and Qian, 2006. Fuller, G. M. and Qian, Y.-Z. (2006). Simultaneous flavor transformation
692
+ of neutrinos and antineutrinos with dominant potentials from neutrino-neutrino forward scatter-
693
+ ing. Phys. Rev. D, 73:023004.
694
+ Gambassi and Silva, 2012. Gambassi, A. and Silva, A. (2012). Large deviations and universality
695
+ in quantum quenches. Phys. Rev. Lett., 109:250602.
696
+ Gaudin, M., 1976. Gaudin, M. (1976). Diagonalisation d’une classe d’hamiltoniens de spin. J.
697
+ Phys. France, 37(10):1087–1098.
698
+ Gorin et al., 2006. Gorin, T., Prosen, T., Seligman, T. H., and ˇZnidariˇc, M. (2006). Dynamics of
699
+ loschmidt echoes and fidelity decay. Physics Reports, 435(2):33–156.
700
+ Grohs et al., 2016. Grohs, E., Fuller, G. M., Kishimoto, C. T., Paris, M. W., and Vlasenko, A.
701
+ (2016). Neutrino energy transport in weak decoupling and big bang nucleosynthesis. Phys. Rev.
702
+ D, 93(8):083522, 1512.02205.
703
+ Hall et al., 2021. Hall, B., Roggero, A., Baroni, A., and Carlson, J. (2021). Simulation of collec-
704
+ tive neutrino oscillations on a quantum computer. Phys. Rev. D, 104(6):063009, 2102.12556.
705
+ Heyl, 2018. Heyl, M. (2018).
706
+ Dynamical quantum phase transitions: a review.
707
+ Reports on
708
+ Progress in Physics, 81(5):054001.
709
+ Heyl et al., 2013. Heyl, M., Polkovnikov, A., and Kehrein, S. (2013). Dynamical quantum phase
710
+ transitions in the transverse-field ising model. Phys. Rev. Lett., 110:135704.
711
+ Husimi, 1940. Husimi, K. (1940). Some formal properties of the density matrix. Proceedings of
712
+ the Physico-Mathematical Society of Japan. 3rd Series, 22(4):264–314.
713
+ ˇZunkoviˇc et al., 2018. ˇZunkoviˇc, B., Heyl, M., Knap, M., and Silva, A. (2018). Dynamical quan-
714
+ tum phase transitions in spin chains with long-range interactions: Merging different concepts of
715
+ nonequilibrium criticality. Phys. Rev. Lett., 120:130601.
716
+ Illa and Savage, 2022. Illa, M. and Savage, M. J. (2022). Basic Elements for Simulations of Stan-
717
+ dard Model Physics with Quantum Annealers: Multigrid and Clock States. 2202.12340.
718
+ Izaguirre et al., 2017. Izaguirre, I., Raffelt, G., and Tamborra, I. (2017). Fast pairwise conversion
719
+ of supernova neutrinos: A dispersion relation approach. Phys. Rev. Lett., 118:021101.
720
+ Janka et al., 2007. Janka, H.-T., Langanke, K., Marek, A., Martinez-Pinedo, G., and Mueller, B.
721
+ (2007). Theory of Core-Collapse Supernovae. Phys. Rept., 442:38–74, astro-ph/0612072.
722
+
723
+ 14
724
+ Amol V. Patwardhan, Michael J. Cervia, Ermal Rrapaj, Pooja Siwach, A. B. Balantekin
725
+ Kajino et al., 2012. Kajino, T., Aoki, W., Cheoun, M. K., Chiba, S., Fujiya, W., Hayakawa, T.,
726
+ Mathews, G. J., Nakamura, K., Shaku, K., and Yoshida, T. (2012). Supernova nucleosynthesis
727
+ and neutrino oscillation. AIP Conf. Proc., 1441(1):375–377.
728
+ Kajino et al., 2014. Kajino, T., Mathews, G. J., and Hayakawa, T. (2014).
729
+ Neutrinos in core-
730
+ collapse supernovae and nucleosynthesis. J. Phys. G, 41:044007.
731
+ Kosteleck´y and Samuel, 1995. Kosteleck´y, V. A. and Samuel, S. (1995). Self-maintained coherent
732
+ oscillations in dense neutrino gases. Phys. Rev. D, 52:621–627.
733
+ Kyutoku et al., 2018. Kyutoku, K., Kiuchi, K., Sekiguchi, Y., Shibata, M., and Taniguchi, K.
734
+ (2018). Neutrino transport in black hole-neutron star binaries: neutrino emission and dynam-
735
+ ical mass ejection. Phys. Rev. D, 97(2):023009, 1710.00827.
736
+ Lacroix and Ayik, 2014. Lacroix, D. and Ayik, S. (2014). Stochastic quantum dynamics beyond
737
+ mean-field. Eur. Phys. J. A, 50:95, 1402.2393.
738
+ Lacroix et al., 2022. Lacroix, D., Balantekin, A. B., Cervia, M. J., Patwardhan, A. V., and Siwach,
739
+ P. (2022). Role of non-gaussian quantum fluctuations in neutrino entanglement. 2205.09384.
740
+ Langanke et al., 2019. Langanke, K., Martinez-Pinedo, G., and Sieverding, A. (2019). Neutrino
741
+ nucleosynthesis: An overview. 1901.03741.
742
+ Martin et al., 2020. Martin, J. D., Carlson, J., and Duan, H. (2020). Spectral swaps in a two-
743
+ dimensional neutrino ring model. Phys. Rev. D, 101:023007.
744
+ Mart´ınez-Pinedo et al., 2017. Mart´ınez-Pinedo, G., Fischer, T., Langanke, K., Lohs, A., Sieverd-
745
+ ing, A., and Wu, M.-R. (2017). Neutrinos and Their Impact on Core-Collapse Supernova Nu-
746
+ cleosynthesis.
747
+ Notzold and Raffelt, 1988. Notzold, D. and Raffelt, G. (1988). Neutrino Dispersion at Finite Tem-
748
+ perature and Density. Nucl. Phys. B, 307:924–936.
749
+ Paeckel et al., 2019. Paeckel, S., K¨ohler, T., Swoboda, A., Manmana, S. R., Schollw¨ock, U.,
750
+ and Hubig, C. (2019). Time-evolution methods for matrix-product states. Annals of Physics,
751
+ 411:167998.
752
+ Pantaleone, 1992a. Pantaleone, J. T. (1992a). Dirac neutrinos in dense matter. Phys. Rev. D,
753
+ 46:510–523.
754
+ Pantaleone, 1992b. Pantaleone, J. T. (1992b). Neutrino oscillations at high densities. Phys. Lett.
755
+ B, 287:128–132.
756
+ Pastor et al., 2002. Pastor, S., Raffelt, G., and Semikoz, D. V. (2002). Physics of synchronized
757
+ neutrino oscillations caused by self-interactions. Phys. Rev. D, 65:053011.
758
+ Patwardhan et al., 2019. Patwardhan, A. V., Cervia, M. J., and Baha Balantekin, A. (2019). Eigen-
759
+ values and eigenstates of the many-body collective neutrino oscillation problem. Phys. Rev. D,
760
+ 99(12):123013, 1905.04386.
761
+ Patwardhan et al., 2021. Patwardhan, A. V., Cervia, M. J., and Balantekin, A. B. (2021). Spec-
762
+ tral splits and entanglement entropy in collective neutrino oscillations.
763
+ Phys. Rev. D,
764
+ 104(12):123035, 2109.08995.
765
+ Pehlivan et al., 2011. Pehlivan, Y., Balantekin, A. B., Kajino, T., and Yoshida, T. (2011). Invari-
766
+ ants of collective neutrino oscillations. Phys. Rev. D, 84:065008.
767
+ Pehlivan et al., 2017. Pehlivan, Y., Subas¸ı, A. L., Ghazanfari, N., Birol, S., and Y¨uksel, H.
768
+ (2017). Spectral splits of neutrinos as a BCS-BEC crossover type phenomenon. Phys. Rev.
769
+ D, 95(6):063022, 1603.06360.
770
+ Qian and Fuller, 1995. Qian, Y. Z. and Fuller, G. M. (1995). Neutrino-neutrino scattering and
771
+ matter enhanced neutrino flavor transformation in Supernovae. Phys. Rev. D, 51:1479–1494,
772
+ astro-ph/9406073.
773
+ Qian et al., 1993. Qian, Y.-Z., Fuller, G. M., Mathews, G. J., Mayle, R., Wilson, J. R., and
774
+ Woosley, S. E. (1993). A Connection between flavor mixing of cosmologically significant neu-
775
+ trinos and heavy element nucleosynthesis in supernovae. Phys. Rev. Lett., 71:1965–1968.
776
+ Raffelt and Smirnov, 2007a. Raffelt, G. G. and Smirnov, A. Y. (2007a). Adiabaticity and spectral
777
+ splits in collective neutrino transformations. Phys. Rev. D, 76:125008, 0709.4641.
778
+ Raffelt and Smirnov, 2007b. Raffelt, G. G. and Smirnov, A. Y. (2007b). Self-induced spectral
779
+ splits in supernova neutrino fluxes. Phys. Rev. D, 76:081301.
780
+ Raffelt and Tamborra, 2010. Raffelt, G. G. and Tamborra, I. (2010). Synchronization versus de-
781
+ coherence of neutrino oscillations at intermediate densities. Phys. Rev. D, 82:125004.
782
+
783
+ Many-body collective neutrino oscillations: recent developments
784
+ 15
785
+ Richardson, 1963. Richardson, R. W. (1963). A restricted class of exact eigenstates of the pairing-
786
+ force Hamiltonian. Physics Letters, 3(6):277–279.
787
+ Richardson, 1965. Richardson, R. W. (1965). Exact Eigenstates of the Pairing-Force Hamiltonian.
788
+ II. Journal of Mathematical Physics, 6(7):1034–1051.
789
+ Richardson and Sherman, 1964. Richardson, R. W. and Sherman, N. (1964). Exact eigenstates of
790
+ the pairing-force Hamiltonian. Nuclear Physics, 52:221–238.
791
+ Richers and Sen, 2022. Richers, S. and Sen, M. (2022). Fast Flavor Transformations. 2207.03561.
792
+ Roberts et al., 2017. Roberts, L. F., Lippuner, J., Duez, M. D., Faber, J. A., Foucart, F., Lombardi,
793
+ J. C., Ning, S., Ott, C. D., and Ponce, M. (2017). The influence of neutrinos on r-process nu-
794
+ cleosynthesis in the ejecta of black hole–neutron star mergers. Mon. Not. Roy. Astron. Soc.,
795
+ 464(4):3907–3919, 1601.07942.
796
+ Roggero, 2021a. Roggero, A. (2021a). Dynamical phase transitions in models of collective neu-
797
+ trino oscillations. Phys. Rev. D, 104:123023.
798
+ Roggero, 2021b. Roggero, A. (2021b). Entanglement and many-body effects in collective neu-
799
+ trino oscillations. Phys. Rev. D, 104:103016.
800
+ Roggero et al., 2022. Roggero, A., Rrapaj, E., and Xiong, Z. (2022). Entanglement and correla-
801
+ tions in fast collective neutrino flavor oscillations. 2203.02783.
802
+ Rrapaj, 2020. Rrapaj, E. (2020).
803
+ Exact solution of multiangle quantum many-body collective
804
+ neutrino-flavor oscillations. Phys. Rev. C, 101:065805.
805
+ Samuel, 1993. Samuel, S. (1993). Neutrino oscillations in dense neutrino gases. Phys. Rev. D,
806
+ 48:1462–1477.
807
+ Sasaki et al., 2017. Sasaki, H., Kajino, T., Takiwaki, T., Hayakawa, T., Balantekin, A. B., and
808
+ Pehlivan, Y. (2017). Possible effects of collective neutrino oscillations in three-flavor multiangle
809
+ simulations of supernova ν p processes. Phys. Rev. D, 96(4):043013, 1707.09111.
810
+ Sawyer, 2004. Sawyer, R. F. (2004). ’Classical’ instabilities and ’quantum’ speed-up in the evo-
811
+ lution of neutrino clouds. hep-ph/0408265.
812
+ Sawyer, 2005. Sawyer, R. F. (2005). Speed-up of neutrino transformations in a supernova envi-
813
+ ronment. Phys. Rev. D, 72:045003, hep-ph/0503013.
814
+ Schollw¨ock, 2011. Schollw¨ock, U. (2011). The density-matrix renormalization group in the age
815
+ of matrix product states. Annals of Physics, 326(1):96–192. January 2011 Special Issue.
816
+ Sciolla and Biroli, 2011. Sciolla, B. and Biroli, G. (2011). Dynamical transitions and quantum
817
+ quenches in mean-field models.
818
+ Journal of Statistical Mechanics: Theory and Experiment,
819
+ 2011(11):P11003.
820
+ Sciolla and Biroli, 2013. Sciolla, B. and Biroli, G. (2013). Quantum quenches, dynamical transi-
821
+ tions, and off-equilibrium quantum criticality. Phys. Rev. B, 88:201110.
822
+ Sigl and Raffelt, 1993. Sigl, G. and Raffelt, G. (1993). General kinetic description of relativistic
823
+ mixed neutrinos. Nucl. Phys. B, 406:423–451.
824
+ Steigman, 2012. Steigman, G. (2012).
825
+ Neutrinos And Big Bang Nucleosynthesis.
826
+ Adv. High
827
+ Energy Phys., 2012:268321, 1208.0032.
828
+ Surman and McLaughlin, 2004. Surman, R. and McLaughlin, G. C. (2004). Neutrinos and nucle-
829
+ osynthesis in gamma-ray burst accretion disks. Astrophys. J., 603:611–623, astro-ph/0308004.
830
+ Tamborra and Shalgar, 2021. Tamborra, I. and Shalgar, S. (2021). New Developments in Flavor
831
+ Evolution of a Dense Neutrino Gas. Ann. Rev. Nucl. Part. Sci., 71:165–188, 2011.01948.
832
+ Vidal, 2003. Vidal, G. (2003). Efficient classical simulation of slightly entangled quantum com-
833
+ putations. Phys. Rev. Lett., 91:147902.
834
+ Volpe et al., 2013. Volpe, C., V¨a¨an¨anen, D., and Espinoza, C. (2013). Extended evolution equa-
835
+ tions for neutrino propagation in astrophysical and cosmological environments. Phys. Rev. D,
836
+ 87(11):113010, 1302.2374.
837
+ Wu et al., 2015. Wu, M.-R., Qian, Y.-Z., Martinez-Pinedo, G., Fischer, T., and Huther, L. (2015).
838
+ Effects of neutrino oscillations on nucleosynthesis and neutrino signals for an 18 M supernova
839
+ model. Phys. Rev. D, 91(6):065016, 1412.8587.
840
+ Xiong, 2022. Xiong, Z. (2022). Many-body effects of collective neutrino oscillations. Phys. Rev.
841
+ D, 105:103002.
842
+
843
+ 16
844
+ Amol V. Patwardhan, Michael J. Cervia, Ermal Rrapaj, Pooja Siwach, A. B. Balantekin
845
+ Xiong et al., 2020. Xiong, Z., Sieverding, A., Sen, M., and Qian, Y.-Z. (2020). Potential Impact
846
+ of Fast Flavor Oscillations on Neutrino-driven Winds and Their Nucleosynthesis. Astrophys. J.,
847
+ 900(2):144, 2006.11414.
848
+ Xiong et al., 2019. Xiong, Z., Wu, M.-R., and Qian, Y.-Z. (2019). Active-sterile Neutrino Oscil-
849
+ lations in Neutrino-driven Winds: Implications for Nucleosynthesis. 1904.09371.
850
+ Yeter-Aydeniz et al., 2022. Yeter-Aydeniz, K., Bangar, S., Siopsis, G., and Pooser, R. C. (2022).
851
+ Collective neutrino oscillations on a quantum computer. Quant. Inf. Proc., 21(3):84, 2104.03273.
852
+ Yoshida et al., 2006. Yoshida, T., Kajino, T., Yokomakura, H., Kimura, K., Takamura, A., and
853
+ Hartmann, D. H. (2006). Supernova neutrino nucleosynthesis of light elements with neutrino
854
+ oscillations. Phys. Rev. Lett., 96:091101, astro-ph/0602195.
855
+
INAyT4oBgHgl3EQfffiq/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
LNE4T4oBgHgl3EQf8A7c/content/tmp_files/2301.05345v1.pdf.txt ADDED
@@ -0,0 +1,1246 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ GOHSP: A Unified Framework of Graph and Optimization-based Heterogeneous
2
+ Structured Pruning for Vision Transformer
3
+ Miao Yin1* Burak Uzkent2, Yilin Shen2, Hongxia Jin2, Bo Yuan1
4
+ 1 Rutgers University, 2 Samsung Research America
5
+ Abstract
6
+ The recently proposed Vision transformers (ViTs) have
7
+ shown very impressive empirical performance in various
8
+ computer vision tasks, and they are viewed as an impor-
9
+ tant type of foundation model. However, ViTs are typically
10
+ constructed with large-scale sizes, which then severely hin-
11
+ der their potential deployment in many practical resources-
12
+ constrained applications. To mitigate this challenging prob-
13
+ lem, structured pruning is a promising solution to compress
14
+ model size and enable practical efficiency. However, unlike
15
+ its current popularity for CNNs and RNNs, structured prun-
16
+ ing for ViT models is little explored.
17
+ In this paper, we propose GOHSP, a unified framework of
18
+ Graph and Optimization-based Structured Pruning for ViT
19
+ models. We first develop a graph-based ranking for measur-
20
+ ing the importance of attention heads, and the extracted im-
21
+ portance information is further integrated to an optimization-
22
+ based procedure to impose the heterogeneous structured spar-
23
+ sity patterns on the ViT models. Experimental results show
24
+ that our proposed GOHSP demonstrates excellent compres-
25
+ sion performance. On CIFAR-10 dataset, our approach can
26
+ bring 40% parameters reduction with no accuracy loss for
27
+ ViT-Small model. On ImageNet dataset, with 30% and 35%
28
+ sparsity ratio for DeiT-Tiny and DeiT-Small models, our ap-
29
+ proach achieves 1.65% and 0.76% accuracy increase over the
30
+ existing structured pruning methods, respectively.
31
+ Introduction
32
+ Recently applying transformer architecture to computer vi-
33
+ sion has emerged as an important forefront of foundation
34
+ model design (Dosovitskiy et al. 2020). Thanks to the del-
35
+ icate vision-specific self-attention, inherent minimal induc-
36
+ tive biases and high scalability and parallelism, vision trans-
37
+ formers (ViTs) (Dosovitskiy et al. 2020; Touvron et al.
38
+ 2021; Zhou et al. 2021) have shown very outstanding and
39
+ even state-of-the-art performance in many fundamental and
40
+ downstream image and video processing tasks, such as im-
41
+ age classification, object detection, super-resolution, video
42
+ classification etc.
43
+ Motivated by the scaling success of the giant natural lan-
44
+ guage processing (NLP) transformers (e.g., BERT (Devlin
45
+ *This work was done during Miao Yin’s internship at Samsung
46
+ Research America.
47
+ Copyright © 2023, Association for the Advancement of Artificial
48
+ Intelligence (www.aaai.org). All rights reserved.
49
+ et al. 2018) and GPT-3 (Brown et al. 2020)), the existing
50
+ ViTs are also constructed with large model sizes to adapt for
51
+ massive data training (Zhai et al. 2021). Consequently, they
52
+ are suffering from huge memory footprints and extensive
53
+ computational costs. These limitations, if not being properly
54
+ addressed, could severely hinder the widespread adoption of
55
+ ViTs in many practical scenarios, especially on the resource-
56
+ constrained mobile platforms and Internet-of-things (IoT)
57
+ devices.
58
+ To mitigate this challenging problem, one attractive solu-
59
+ tion is to perform model compression (Yu et al. 2017; Kim
60
+ et al. 2016; Pan et al. 2019) to reduce the network costs with-
61
+ out affecting task performance. However, unlike the current
62
+ popularity of compressing convolutional and recurrent neu-
63
+ ral networks (CNNs and RNNs), ViT-oriented model com-
64
+ pression has not been systematically studied yet. In partic-
65
+ ular, structured pruning, as an important hardware-friendly
66
+ compression strategy that can bring practical efficiency on
67
+ the off-the-shelf hardware, is little explored for ViT models.
68
+ To date, a rich set of structured pruning approaches have
69
+ been proposed and investigated in the existing literatures,
70
+ and most of them focus on sparsifying the CNNs at the chan-
71
+ nel level (He, Zhang, and Sun 2017; Ye et al. 2018). On the
72
+ other hand, as will be analyzed and elaborated in Section ,
73
+ because of the difference of the underlying architecture, the
74
+ structured sparse ViT models can exhibit multi-granularity
75
+ sparsity (i.e., head-level and column-level) in the different
76
+ component modules (i.e., attention head and multi-layer per-
77
+ ception (MLP)). The co-existence of such heterogeneous
78
+ sparse patterns raises a series of new research challenges and
79
+ questions when we consider the efficient structured pruning
80
+ strategy for ViT models. For instance, for each component
81
+ module what is the corresponding suitable pruning criterion
82
+ to obtain its specific sparse pattern? Also, how should we
83
+ perform the entire pruning process across different modules
84
+ with different levels of granularity sparsity to optimize the
85
+ overall compression and task performance?
86
+ Technical Preview & Contributions. To answer these
87
+ questions, in this paper we propose GOHSP, a unified frame-
88
+ work of Graph and Optimization-based Structure Prun-
89
+ ing for vision transformer. To be specific, we first de-
90
+ velop a graph-based ranking approach to measure the im-
91
+ portance of attention heads. As a soft-pruning guideline,
92
+ such importance information is then integrated to the overall
93
+ arXiv:2301.05345v1 [cs.AI] 13 Jan 2023
94
+
95
+ optimization-based procedure to impose the different types
96
+ of structured sparsity in a joint and global way. Overall, the
97
+ contributions of this paper are summarized as follows:
98
+ • We propose a graph-based ranking algorithm to mea-
99
+ sure and determine the importance of attention heads.
100
+ By modeling the inter-head correlation as a converged
101
+ Markov chain, the head importance can be interpreted
102
+ and calculated as the stationary distribution, which is fur-
103
+ ther used as a soft guideline for the overall pruning pro-
104
+ cedure.
105
+ • We propose a unified framework to jointly optimize dif-
106
+ ferent types of structured sparsity in the different mod-
107
+ ules. The complicated coordination for different sparse
108
+ patterns are automatically learned and optimized in a sys-
109
+ tematic way.
110
+ • We evaluate the performance of our structured pruning
111
+ approach of different ViT models on different datasets.
112
+ On CIFAR-10 dataset, our approach can bring 40% pa-
113
+ rameters reduction with no accuracy loss for ViT-Small
114
+ model. On ImageNet dataset, with 30% and 40% spar-
115
+ sity ratio for DeiT-Tiny and DeiT-Small models, our
116
+ approach achieves 1.65% and 0.76% accuracy increase
117
+ than the existing structured pruning methods, respec-
118
+ tively.
119
+ Related Work
120
+ Vision Transformer. Inspired by the grand success of trans-
121
+ former architecture in NLP domains, deep learning re-
122
+ searchers have actively explored the efficient transformer-
123
+ based neural networks for computer vision. Most recently,
124
+ several vision transformers (ViTs) and their variants have
125
+ already shown very impressive performance in several im-
126
+ age and video processing tasks (Dosovitskiy et al. 2020;
127
+ Touvron et al. 2021; Zhou et al. 2021). However, in order
128
+ to achieve competitive performance with the state-of-the-art
129
+ CNNs, ViTs typically have to scale up their model sizes and
130
+ therefore they suffer from costly computation and storage.
131
+ Dynamic Inference with ViTs. To reduce the deploy-
132
+ ment costs of ViTs, several works (Wang et al. 2021;
133
+ Bakhtiarnia, Zhang, and Iosifidis 2021; Rao et al. 2021;
134
+ Meng et al. 2022; Xu et al. 2022; Uzkent, Yeh, and Er-
135
+ mon 2020; Uzkent and Ermon 2020) have been proposed to
136
+ improve the processing speed via dynamically pruning the
137
+ tokens/patches or skipping transformer components adap-
138
+ tively. Essentially as dynamic inference approaches, this set
139
+ of works do not pursue to reduce the model sizes but focus
140
+ on input-aware inference to obtain practical speedup. Our
141
+ structured pruning-based solution is orthogonal to them, and
142
+ these two different strategies can be potentially combined
143
+ together to achieve higher speed and smaller memory foot-
144
+ print.
145
+ Structured Pruning. Model compression is a promising
146
+ strategy to reduce the deployment costs of neural networks.
147
+ Among various model compression techniques, structured
148
+ pruning is a very popular choice because its hardware-
149
+ friendly nature can bring practical efficiency on the real-
150
+ world devices. Based on different pruning criterias, various
151
+ structured pruning approaches have been extensively studied
152
+ Head-based Sparsity
153
+ (Multi-Head Attention)
154
+ (a) Conventional Unstructured Sparsity
155
+ (b) Unified Structured Sparsity (Ours)
156
+ Column-based Sparsity
157
+ (Multi-Head Attention)
158
+ Column-based Sparsity
159
+ (MLP)
160
+ Unstructured Sparsity
161
+ (MLP)
162
+ Unstructured Sparsity
163
+ (Multi-Head Attention)
164
+ Figure 1: (a) Sparsity pattern of ViT models after un-
165
+ structured pruning. Only part of the Multi-Head Attention
166
+ and MLP columns are pruned which are not hardware-
167
+ friendly.(b) Heterogeneous sparsity patterns of ViT models
168
+ after structured pruning. Certain MLP and Multi-Head At-
169
+ tention columns are removed which is hardware-friendly.
170
+ in the existing literature (Yu et al. 2018; Zhuang et al. 2018;
171
+ Liu et al. 2019; He et al. 2019; Lin et al. 2020; Tiwari et al.
172
+ 2021; Lou et al. 2022), and most of them focus on pruning
173
+ CNN models; while the efficient structured pruning of ViTs
174
+ is little explored. One of these studies, (Chen et al. 2021),
175
+ prunes the vision transformers using structured pruning. (Yu
176
+ et al. 2022), on the other hand, focuses on FLOPs reduction
177
+ with the vision transformers using pruning, layer skipping,
178
+ and knowledge distillation whereas in our study we focus on
179
+ structured pruning to mainly reduce the number of parame-
180
+ ters for building hardware-friendly compressed models. For
181
+ this reason, we compare our method to (Chen et al. 2021).
182
+ Structured Pruning of ViTs: Analysis
183
+ Notation. Considering an L-block vision transformer,
184
+ W (l)
185
+ attn = {W (l)
186
+ qkv, W (l)
187
+ proj} and W (l)
188
+ mlp = {W (l)
189
+ fc1, W (l)
190
+ fc2} rep-
191
+ resent the weights of the attention layer and the MLP layer
192
+ at l-th block, respectively. For each attention layer, there are
193
+ H self-attention heads, namely W (l)
194
+ qkv = {W (l,h)
195
+ qkv }H
196
+ h=1 and
197
+ W (l)
198
+ proj = {W (l,h)
199
+ proj }H
200
+ h=1. To simplify the notation, in the fol-
201
+ lowing content we take one block as the example and omit
202
+ the superscript (layer index).
203
+ Heterogeneity of structured sparsity. Because of the
204
+ difference of the network architecture, the meaning of
205
+ ‘structured sparsity’ varies with different model types. As
206
+ described and performed in (Wen et al. 2016; Anwar,
207
+ Hwang, and Sung 2017; Liu et al. 2018, 2020), the struc-
208
+ tured pruning of CNN and RNN typically indicates the re-
209
+ moval of the entire channels of the weight tensors and the
210
+ entire columns of the weight matrices, respectively. Notice
211
+ that here for either of these two cases, only one type of the
212
+ structured sparse pattern exist because of the architectural
213
+ homogeneity of the CNN and RNN.
214
+ On the other hand, a ViT model exhibits inherent archi-
215
+ tectural heterogeneity. Within the same block, the front-end
216
+ multi-head attention module and the back-end MLP mod-
217
+ ule represent two types of design philosophy for information
218
+ processing, and thereby leading to huge difference on both
219
+ computing procedures and the available structured sparse
220
+ patterns.
221
+ To be specific, when we consider performing structured
222
+
223
+ pruning of ViT model, three types of structured sparse pat-
224
+ terns can co-exist with different levels of granularity across
225
+ different modules. For the multi-head attention module, be-
226
+ cause each attention head is processing the information in-
227
+ dividually in a parallel way, the pruning can be performed
228
+ at the head-level to sparsify this component. In addition,
229
+ consider the weights in the heads are represented in the ma-
230
+ trix format; the column-level sparsity can also be introduced
231
+ towards structured pruning. Meanwhile, because the MLP
232
+ consists of multiple weight matrices as well, the column-
233
+ level of granularity sparsity can be imposed on this back-end
234
+ module at the same time. Consequently, a structured pruned
235
+ ViT model can exhibit heterogeneous structured sparsity
236
+ (see Fig. 1).
237
+ Problem Definition. Based on the above analysis, the
238
+ structured pruning of a vision transformer model with loss
239
+ function ℓ(·) can be formulated as the following general op-
240
+ timization problem:
241
+ min
242
+ Wattn,Wmlpℓ(Wattn, Wmlp),
243
+ s.t.
244
+ ∥Wattn∥h
245
+ 0 ≤ κh
246
+ attn,
247
+ ∥Wattn∥c
248
+ 0 ≤ κc
249
+ attn,
250
+ ∥Wmlp∥c
251
+ 0 ≤ κc
252
+ mlp,
253
+ (1)
254
+ where κc and κh are the desired number of columns and the
255
+ desired number of heads after pruning, respectively. ∥ · ∥c
256
+ 0
257
+ and ∥ · ∥h
258
+ 0 are the column-based and head-based group L0-
259
+ norm, which denote the number of non-zero columns and
260
+ the number of non-zero heads, respectively.
261
+ Questions to be Answered. Solving the above opti-
262
+ mization problem is non-trivial since it contains the con-
263
+ straints involved with multi-granularity sparsity for different
264
+ model components. More specifically, two important ques-
265
+ tions need to be answered. Question #1: What is the suitable
266
+ pruning criterion to obtain head-level sparsity?
267
+ Analysis: From the perspective of information process-
268
+ ing, multi-head attention shares some interesting similarity
269
+ with convolutional layer. Both of them use multiple indi-
270
+ vidual computing units, i.e., attention heads and convolu-
271
+ tional filters, to perform parallel computations. Therefore, a
272
+ naive way to perform head-level pruning is to leverage the
273
+ existing criteria developed in the channel pruning of CNNs.
274
+ However, such straightforward solution, in principle, may
275
+ not be the best choice because of two reasons. First, the re-
276
+ ceptive fields and the focused locality of the attention head
277
+ and filters are different, and hence simply using the crite-
278
+ rion for pruning channels is not a suitable strategy. Second
279
+ and more importantly, most of the existing channel pruning
280
+ criterias are built on the information of each individual chan-
281
+ nel (the corresponding filter weight and/or its feature map).
282
+ When adopting this philosophy in the head pruning, the in-
283
+ sufficient utilization of inter-head information will probably
284
+ cause non-negligible performance loss. Overall, the unique
285
+ characteristics of multi-head attention mechanism calls for
286
+ attention-specific pruning criterion.
287
+ Question #2: How should we coordinate the pruning
288
+ across different modules with different levels of granularity?
289
+ Analysis: As indicated before, three types of structured
290
+ sparse pattern can co-exist in the different modules of the
291
+ pruned ViT models. A key component of the to-be-explored
292
+ structured pruning strategy is to develop a good coordination
293
+ scheme that can properly impose these different structured
294
+ sparse patterns in a joint and global way. Consider the com-
295
+ plicated interaction among different types of structured spar-
296
+ sity, the expected pruning strategy should be able to solve
297
+ this problem in a systematic and global way.
298
+ Structured Pruning of ViTs: Method
299
+ Graph-based Head Ranking
300
+ To answer Question #1, we propose a graph-based approach
301
+ to measure and determine the importance of different at-
302
+ tention heads, which can be further used for the follow-up
303
+ pruning. Our key idea is to model the inter-head correla-
304
+ tion as a graph, and then leverage the graph-based ranking,
305
+ a methodology that has been successfully used in many web
306
+ search and NLP algorithms, such as PageRank (Page et al.
307
+ 1999), TextRank (Mihalcea and Tarau 2004) and LexRank
308
+ (Erkan and Radev 2004), to select important attention heads.
309
+ Graph Construction of Markov Chain. To be specific,
310
+ we first construct a graph G = (A, E) to represent the atten-
311
+ tion heads and their similarities in the block of a ViT model.
312
+ The set of nodes A denote all the attention heads {Ah}H
313
+ h=1,
314
+ and E is the set of connected edges. For edge E(Ai, Aj),
315
+ its weight is defined as the expected cosine similarity be-
316
+ tween Ai and Aj. According to (Mihalcea and Tarau 2004),
317
+ the graph defined with such cosine similarity can be inter-
318
+ preted as a Markov chain, where each node is a state, and
319
+ the transition probability P(i, j) between two states is the
320
+ edge weight. In such scenario, P(i, j) can be calculated as:
321
+ P(i, j) = EX∼D [CosineSim(Ai(X), Aj(X))] ,
322
+ (2)
323
+ where Ai(X) is the output of i-th attention head with sam-
324
+ pled input X and D is the data set. Built upon this calcula-
325
+ tion, the entire transition matrix P of a Markov chain. No-
326
+ tice that as indicated in (Erkan and Radev 2004), each col-
327
+ umn of P should be further normalized.
328
+ Batch estimation. Calculating the transition probability
329
+ can be very costly since it needs to be performed across the
330
+ entire training dataset D (see Eq. 2). To solve this problem,
331
+ we adopt a batch-based estimation strategy to improve com-
332
+ putation efficiency without sacrificing ranking performance.
333
+ To be specific, as described in Eq. 3, only a batch of training
334
+ data is sampled and used to to estimate the transition prob-
335
+ ability. As our ablation study in Section will show, using
336
+ different batch sizes (B) bring very stable ranking results
337
+ for the attention heads, thereby empirically verifying the ef-
338
+ fectiveness of this estimation strategy.
339
+ P(i, j) = CosineSim
340
+ � B
341
+
342
+ b=1
343
+ Ai(Xb),
344
+ B
345
+
346
+ b=1
347
+ Aj(Xb)
348
+
349
+ .
350
+ (3)
351
+ Importance Ranking. Mathematically, an irreducible
352
+ and aperiodic Markov chain is guaranteed to converge to a
353
+ stationary distribution (Seneta 2006). As indicated in (Erkan
354
+ and Radev 2004), once converged, the probability of a ran-
355
+ dom walker stays in one state can reflect the state impor-
356
+ tance. Motivated by this observation, we propose to quantify
357
+
358
+ Multi-Head Attention
359
+ Embedded
360
+ Patches
361
+ MLP
362
+ Block
363
+ 0.1
364
+ Graph-based Heads Ranking
365
+ Multi-Head Attention
366
+ Embedded
367
+ Patches
368
+ MLP
369
+ Block
370
+ Data
371
+ Optimization-based Soft Pruning
372
+ Multi-Head Attention
373
+ Embedded
374
+ Patches
375
+ MLP
376
+ Block
377
+ Fine-Tuning
378
+ X
379
+ Score
380
+ Mask
381
+ 0.7
382
+ 0.5
383
+ 0.3
384
+ Normalize
385
+ Figure 2: Procedure of the proposed multi-stage structured pruning approach.
386
+ the importance of each attention head via calculating the sta-
387
+ tionary distribution in our constructed Markov chain. To that
388
+ end, the iterative power method (Erkan and Radev 2004) can
389
+ be used via setting a uniform distribution for the states as the
390
+ initialization. Overall, the entire graph-based head ranking
391
+ procedure is described in Algorithm 1.
392
+ Soft-Pruning Mask. Once the importance score for each
393
+ state is obtained via calculating the stationary distribution,
394
+ the corresponding attention heads can be ranked. Here we
395
+ use a binary mark matrix Mattn = {Mqkv, Mproj} to in-
396
+ dicate the weight entries associated with the least important
397
+ heads that should be removed. Notice that at this stage the
398
+ head pruning is not performed yet. Instead such binary mask
399
+ serves as the guideline for the next-stage optimization, and
400
+ it is essentially integrated into Eq. 1 as follows:
401
+ min
402
+ Wattn,Wmlpℓ(Wattn, Wmlp)
403
+ s.t.
404
+ ∥(1 − Mattn) ⊙ Wattn∥0 = 0,
405
+ ∥Wmlp∥0 ≤ κc
406
+ mlp,
407
+ ∥Mattn ⊙ Wattn∥c
408
+ 0 ≤ κc
409
+ attn,
410
+ (4)
411
+ where ⊙ is element-wise product. In general, because the
412
+ overall optimization phase coordinates and adjusts the dif-
413
+ ferent types of structured sparse pattern from a global per-
414
+ Algorithm 1: Graph-based Attention Head Ranking
415
+ Input: Sampled batch {Xb}B
416
+ b=1, attention heads {Ah}H
417
+ h=1;
418
+ Output: Importance score s = [s1, · · · , sH].
419
+ 1: Initialize transition matrix: P := zeros(H, H);
420
+ 2: for i = 1 to H do
421
+ 3:
422
+ for j = 1 to H do
423
+ 4:
424
+ Calculate P(i, j) via Eq. 3;
425
+ 5: Normalize each column of P ;
426
+ 6: Initialize s := ones(H)/H;
427
+ 7: repeat
428
+ 8:
429
+ s′ := s;
430
+ 9:
431
+ s := P s;
432
+ 10:
433
+ δ := ∥s − s′∥;
434
+ 11: until δ ≤ ϵ
435
+ spective, this ranking-only ”soft” pruning strategy, instead
436
+ of directly pruning the least important heads, can provide
437
+ more flexibility and possibility for the next-stage optimiza-
438
+ tion procedure to identify better structured sparse models.
439
+ Optimization-based Structured Pruning
440
+ As pointed out by Question #2, the co-existence of multi-
441
+ granularity and multi-location of the sparsity of ViT models
442
+ make the entire structured pruning procedure become very
443
+ challenging. To solve this, we propose to use advanced op-
444
+ timization technique to perform systematic structured prun-
445
+ ing. To be specific, considering the complicated interactions
446
+ among different types of structured sparsity, we do not prune
447
+ the heads or columns immediately, since any direct hard
448
+ pruning at the early stage may cause severe accuracy loss.
449
+ Instead, we adopt ”soft-pruning” strategy via optimizing the
450
+ entire ViT models towards the desired structured sparse for-
451
+ mats. In other words, the three types of sparsity pattern are
452
+ gradually imposed onto the attention heads and MLPs.
453
+ To that end, we first relax the constraints of Eq. 4 and
454
+ rewrite it as follows:
455
+ min
456
+ Wattn,Wmlpℓ(Wattn, Wmlp) + λ
457
+ 2 ∥(1 − Mattn) ⊙ Wattn∥2
458
+ F ,
459
+ s.t.
460
+ ∥Wmlp∥c
461
+ 0 ≤ κc
462
+ mlp,
463
+ ∥Mattn ⊙ Wattn∥c
464
+ 0 ≤ κc
465
+ attn,
466
+ (5)
467
+ where λ is the coefficient that controls the influence of
468
+ quadratic term.
469
+ Optimization-based Soft Pruning. As indicated in
470
+ (Boyd, Parikh, and Chu 2011), when the constraints of con-
471
+ tinuous non-convex problem are sparsity related (as Eq.
472
+ 5 shows), Douglas—Rachford splitting method (Eckstein
473
+ and Bertsekas 1992) can be a suitable optimization solution
474
+ for such types of problem. Following this philosophy, we
475
+ first introduce auxiliary variables Zattn, Zmlp and indicator
476
+ functions as:
477
+ g(Zattn) =
478
+ �0
479
+ ∥Mattn ⊙ Zattn∥c
480
+ 0 ≤ κc
481
+ attn,
482
+ +∞
483
+ otherwise,
484
+ (6)
485
+ h(Zmlp) =
486
+ �0
487
+ ∥Zmlp∥c
488
+ 0 ≤ κc
489
+ mlp,
490
+ +∞
491
+ otherwise.
492
+ (7)
493
+
494
+ Algorithm 2: Overall Procedure of GOHSP Framework
495
+ Input: Dense weight {Wattn, Wmlp}, desired model size
496
+ {κattn, κmlp}, training data D, number of epochs E;
497
+ Output: Structured sparse weight { ˜
498
+ Wattn, ˜
499
+ Wmlp};
500
+ 1: Sample a batch of data {Xb}B
501
+ b=1 from D;
502
+ 2: Calculate importance score s via Alg. 1;
503
+ 3: Obtain structured mask Mattn according to s;
504
+ 4: Zattn := Wattn, Zmlp := Wmlp; // Initialize auxiliary
505
+ variables
506
+ 5: Uattn := 0, Umlp := 0; // Initialize Lagrangian multi-
507
+ pliers
508
+ 6: for e = 1 to E do
509
+ 7:
510
+ Update Wattn, Wattn via Eq. 10 and Eq. 11;
511
+ 8:
512
+ Update Zattn, Zmlp via Eq. 12 and Eq. 13;
513
+ 9:
514
+ Update Uattn, Umlp via Eq. 14 and Eq. 15;
515
+ 10: Fine-tune pruned weight { ˜
516
+ Wattn, ˜
517
+ Wmlp}.
518
+ Then, we can rewrite Eq. 5 as the following equivalent form:
519
+ min
520
+ W ,Z
521
+ ℓ(Wattn, Wmlp) + g(Zattn) + h(Zmlp)+
522
+ λ
523
+ 2 ∥(1 − Mattn) ⊙ Wattn∥2
524
+ F ,
525
+ s.t.
526
+ Wmlp = Zmlp,
527
+ Wattn = Zattn.
528
+ (8)
529
+ In such scenario, the corresponding augmented Lagrangian
530
+ function of the above optimization objective is:
531
+ Lρ(Wattn, Wmlp, Zmlp) = ℓ(Wattn, Wmlp) + g(Zattn)+
532
+ h(Zmlp) + λ
533
+ 2 ∥(1 − Mattn) ⊙ Wattn∥2
534
+ F +
535
+ ρ
536
+ 2∥Wattn − Zattn + Uattn∥2
537
+ F +
538
+ ρ
539
+ 2∥Uattn∥2
540
+ F + ρ
541
+ 2∥Wmlp − Zmlp + Umlp∥2
542
+ F + ρ
543
+ 2∥Umlp∥2
544
+ F ,
545
+ (9)
546
+ where ρ > 0 is the penalty parameter, and Uattn, Umlp are
547
+ the Lagrangian multipliers. Then the variables at step t can
548
+ be iteratively updated as:
549
+ W t
550
+ attn = W t−1
551
+ attn−η
552
+ ℓ(Wattn, W t−1
553
+ mlp )
554
+ Wattn
555
+
556
+ λ
557
+
558
+ (1 − Mattn) ⊙ W t−1
559
+ attn
560
+
561
+ −ρ(W t−1
562
+ attn − Zt−1
563
+ attn + U t−1
564
+ attn),
565
+ (10)
566
+ W t
567
+ mlp = W t−1
568
+ mlp − η ℓ(W t
569
+ attn, Wmlp)
570
+ Wmlp
571
+
572
+ ρ(W t−1
573
+ mlp − Zt−1
574
+ mlp + U t−1
575
+ mlp ),
576
+ (11)
577
+ Zt
578
+ attn = P(W t
579
+ attn + U t−1
580
+ attn),
581
+ (12)
582
+ Zt
583
+ mlp = P(W t
584
+ mlp + U t−1
585
+ mlp ),
586
+ (13)
587
+ U t
588
+ attn = U t−1
589
+ attn + W t
590
+ attn − Zt
591
+ attn,
592
+ (14)
593
+ U t
594
+ mlp = U t−1
595
+ mlp + W t
596
+ mlp − Zt
597
+ mlp.
598
+ (15)
599
+ Here η is the optimizer learning rate for training the ViT, and
600
+ P is the Euclidean projection for the sparse constraint.
601
+ Final Hard-Pruning and Fine-Tuning. After the above
602
+ described optimization procedure, the structured sparse pat-
603
+ terns have been gradually imposed onto the ViT model.
604
+ In other words, the weight values of the masked attention
605
+ heads, as well as some columns of MLPs and attention
606
+ heads, become extremely small. At this stage, we can now
607
+ prune those small weights and then perform a few rounds of
608
+ fine-tuning to achieve higher performance.
609
+ Overall,
610
+ by
611
+ using
612
+ graph-based
613
+ head
614
+ ranking
615
+ and
616
+ optimization-based structured pruning, the previously raised
617
+ Question #1 and #2 can be properly addressed. The overall
618
+ GOHSP framework is summarized in Fig. 2.
619
+ Experiments
620
+ Experimental Settings
621
+ Dataset and Baseline. We evaluate the performance of
622
+ our proposed GOHSP approach on CIFAR-10 and Ima-
623
+ geNet datasets (Deng et al. 2009). For experiments on
624
+ the CIFAR-10 dataset, the original dense model is ViT-
625
+ Small1(Dosovitskiy et al. 2020) with 48M parameters. For
626
+ experiments on the ImageNet dataset, the original dense
627
+ models are DeiT-Tiny and DeiT-Small (Touvron et al. 2021)
628
+ with 5.7M and 22.1M parameters, respectively.
629
+ Hyper-parameters and Sparsity Ratio. For our experi-
630
+ ments on the CIFAR-10 dataset, the batch size, learning rate
631
+ and ρ are set as 256, 0.1 and 0.001, respectively. For Ima-
632
+ geNet dataset, the batch size, learning rate and ρ are set as
633
+ 256, 0.01 and 0.001, respectively. For both of these two ex-
634
+ periments, SGD is selected as the training optimizer with-
635
+ out using weight decay, and we apply Erd˝os-R´enyi (Mo-
636
+ canu et al. 2018) to determine the sparsity distribution of
637
+ each layer given an overall sparsity ratio. In particular, soft-
638
+ pruning maintains high accuracy at the high sparsity ratios.
639
+ Performance Evaluation
640
+ CIFAR-10 Dataset. Table 1 shows performance compari-
641
+ son on CIFAR-10 dataset between our proposed GOHSP
642
+ and other structured pruning method (structured one-shot
643
+ magnitude pruning (SOMP) (Han, Mao, and Dally 2015)
644
+ and structured gradually magnitude pruning (SGMP) (Zhu
645
+ and Gupta 2017)) for ViT-Small model. It is seen that
646
+ with the same sparsity ratio, our approach brings significant
647
+ performance improvement. Compared to SGMP approach,
648
+ 1We take this model from open source library timm.
649
+ Table 1: Performance comparison between our GOHSP
650
+ and structured one-shot/gradually magnitude-based pruning
651
+ (SOMP/SGMP) of ViT-Small model on CIFAR-10 dataset.
652
+ Method
653
+ Sparsity
654
+ # Paramters
655
+ Top-1 (%)
656
+ Baseline
657
+ -
658
+ 48.0M
659
+ 97.85
660
+ SOMP
661
+ 40%
662
+ 28.8M
663
+ 96.07
664
+ SGMP
665
+ 40%
666
+ 28.8M
667
+ 96.93
668
+ GOHSP (Ours)
669
+ 40%
670
+ 28.8M
671
+ 97.89
672
+ GOHSP (Ours)
673
+ 80%
674
+ 9.6M
675
+ 97.40
676
+
677
+ Table 2: Comparison results of our method, GOHSP, with other structured and unstructured pruning methods on ImageNet.
678
+ Model
679
+ Method
680
+ Sparsity
681
+ # Parameters
682
+ FLOPs ↓
683
+ Run-time ↓
684
+ Top-1 (%)
685
+ DeiT-Tiny
686
+ Baseline
687
+ -
688
+ 5.7M
689
+ -
690
+ -
691
+ 72.20
692
+ OMP (Unstructured)
693
+ 30%
694
+ 4.02M
695
+ 25.56%
696
+ -
697
+ 68.35
698
+ GMP (Unstructured)
699
+ 30%
700
+ 4.02M
701
+ 25.56%
702
+ -
703
+ 69.56
704
+ TP (Unstructured)
705
+ 30%
706
+ 4.02M
707
+ 25.56%
708
+ -
709
+ 68.38
710
+ SSP (Structured)
711
+ 30%
712
+ 4.2M
713
+ 23.69%
714
+ -
715
+ 68.59
716
+ S2ViTE (Structured)
717
+ 30%
718
+ 4.2M
719
+ 23.69%
720
+ 10.57 %
721
+ 70.12
722
+ GOHSP (Structured)
723
+ 30%
724
+ 4.0M
725
+ 30%
726
+ 13.41%
727
+ 70.24
728
+ DeiT-Small
729
+ Baseline
730
+ -
731
+ 22.1M
732
+ -
733
+ -
734
+ 79.90
735
+ SSP (Structured)
736
+ 40%
737
+ 14.6M
738
+ 31.63%
739
+ -
740
+ 77.74
741
+ S2ViTE (Structured)
742
+ 40%
743
+ 14.6M
744
+ 31.63%
745
+ 22.65%
746
+ 79.22
747
+ GOHSP (Structured)
748
+ 40%
749
+ 14.4M
750
+ 35%
751
+ 24.61%
752
+ 79.98
753
+ GOHSP (Structured)
754
+ 50%
755
+ 11.1M
756
+ 39%
757
+ 26.57%
758
+ 79.86
759
+ GOHSP achieves 0.96% accuracy increase with the same
760
+ pruned model size. Even compared with the baseline, the
761
+ structured sparse model pruned by GOHSP can outperform
762
+ the uncompressed model with 40% fewer parameters while
763
+ 80% compressed model achieves only 0.45% worse than the
764
+ full ViT-Small model.
765
+ ImageNet Dataset. Table 2 summarizes the performance
766
+ on ImageNet dataset between GOHSP and other structured
767
+ pruning approaches (SOMP, SGMP, Talyer pruning (TP),
768
+ Salience-based Structured Pruning (SSP) and S2ViTE(Chen
769
+ et al. 2021)) for DeiT-Tiny and DeiT-Small models. It is seen
770
+ that due to the limited redundancy in such small-size model,
771
+ the existing pruning approaches suffer from more than 2.5%
772
+ accuracy loss when compressing DeiT-Tiny. Instead, with
773
+ the even fewer parameters and more FLOPs reduction, our
774
+ GOHSP approach can achieve at least 0.68% accuracy in-
775
+ crease over the unstructured pruning approaches. Compared
776
+ to the structured pruning approach (SSP), our method enjoys
777
+ 1.65% accuracy improvement with lower storage cost and
778
+ computational cost. In addition, when compressing DeiT-
779
+ Small model, with fewer parameters and more FLOPs re-
780
+ duction, our GOHSP approach can achieve 0.76% accuracy
781
+ increase as compared to the state-of-the-art structured prun-
782
+ ing method S2ViTE (Chen et al. 2021) and can even outper-
783
+ form the original DeiT-Small. With 50% pruned DeiT-Small
784
+ we achieve similar accuracy to the full DeiT-Small. Finally,
785
+ we report 26.57% improvement in run-time efficiency with
786
+ our 50% pruned DeiT-Small.
787
+ 0.2
788
+ 0.3
789
+ 0.4
790
+ 0.5
791
+ 0.6
792
+ Sparsity
793
+ 90.0
794
+ 92.0
795
+ 94.0
796
+ 96.0
797
+ 98.0
798
+ Top-1 Accuracy (%)
799
+ Ours
800
+ Hard Pruning
801
+ Figure 3: Results on the effect of soft-pruning (ours) and
802
+ hard-pruning for ViT-Small model on CIFAR-10 dataset.
803
+ Ablation Study, Visualization and Discussion
804
+ To obtain the deep understanding of the effect of our pro-
805
+ posed approach, we perform several ablation studies and a
806
+ detailed analysis. Here the experiments conducted in the ab-
807
+ lation study focus on compressing ViT-Small on CIFAR-10.
808
+ Soft Pruning vs Hard Pruning. As described in Opti-
809
+ mization section, after ranking the attention heads, we use
810
+ the ranking information as a soft-pruning mask to guide
811
+ the next-phase optimization. The optimization itself is also
812
+ a soft-pruning procedure that does not directly zero the
813
+ weights but gradually impose the structured sparsity. To ana-
814
+ lyze the effect of this strategy, we conduct an ablation exper-
815
+ iment via performing the direct hard pruning. In this ablation
816
+ study, the least important attention heads are removed ac-
817
+ cording to their ranks, and the columns of MLPs with least
818
+ group L1 norm are also pruned. Such hard pruned models
819
+ are still trained with the same hyper-parameters settings that
820
+ are used for soft pruning method. Fig. 3 shows the curves
821
+ of top-1 test accuracy with different target sparsity settings.
822
+ 1
823
+ 2
824
+ 3
825
+ 4
826
+ 5
827
+ 6
828
+ 7
829
+ 8
830
+ Head Index
831
+ 1
832
+ 2
833
+ 3
834
+ 4
835
+ 5
836
+ 6
837
+ 7
838
+ 8
839
+ Block Index
840
+ Batch Size=256
841
+ 0
842
+ 2
843
+ 4
844
+ 6
845
+ 1
846
+ 2
847
+ 3
848
+ 4
849
+ 5
850
+ 6
851
+ 7
852
+ 8
853
+ Head Index
854
+ 1
855
+ 2
856
+ 3
857
+ 4
858
+ 5
859
+ 6
860
+ 7
861
+ 8
862
+ Block Index
863
+ Batch Size=512
864
+ 0
865
+ 2
866
+ 4
867
+ 6
868
+ 1
869
+ 2
870
+ 3
871
+ 4
872
+ 5
873
+ 6
874
+ 7
875
+ 8
876
+ Head Index
877
+ 1
878
+ 2
879
+ 3
880
+ 4
881
+ 5
882
+ 6
883
+ 7
884
+ 8
885
+ Block Index
886
+ Batch Size=1024
887
+ 0
888
+ 2
889
+ 4
890
+ 6
891
+ 1
892
+ 2
893
+ 3
894
+ 4
895
+ 5
896
+ 6
897
+ 7
898
+ 8
899
+ Head Index
900
+ 1
901
+ 2
902
+ 3
903
+ 4
904
+ 5
905
+ 6
906
+ 7
907
+ 8
908
+ Block Index
909
+ Batch Size=1536
910
+ 0
911
+ 2
912
+ 4
913
+ 6
914
+ Figure 4: The effect of batch sizes for ranking results. Dif-
915
+ ferent colors represent different ranking scores. We can see
916
+ that our head ranking algorithm is not sensitive to batch size.
917
+
918
+ The soft-pruning strategy brings very significant accuracy
919
+ improvement over the direct hard pruning with the same
920
+ sparsity ratio.
921
+ Effect of Batch Size on Head Ranking. As shown in Eq.
922
+ 3, the importance scores of attention head is calculated on
923
+ a batch of data. To investigate the potential impact of batch
924
+ sizes for the ranking results, we observe the change of rank-
925
+ ing with different batch sizes. As shown in Fig. 4, the rank-
926
+ ing results are very stable (almost the same) when the batch
927
+ size varies. Therefore we can conclude that using batches of
928
+ data can already achieve very good estimation of head rank-
929
+ ing. In other words, our ranking approach has low sensitivity
930
+ to the distribution of input data.
931
+ Sensitivity of Penalty Parameter ρ. We also explore the
932
+ effect of hyperparameter ρ on the structured pruning proce-
933
+ dure. Fig. 5 (a) shows the convergence of training process
934
+ with respect to different ρ. It is seen that the convergence
935
+ speed is always fast, and hence it demonstrates the promis-
936
+ ing convergence property of our approach in practice. Fig. 5
937
+ (b) illustrates the L2-norm of the masked entries. It is seen
938
+ that the larger ρ makes the model exhibit more sparsity at the
939
+ earlier stage, thereby indicating that larger ρ can bring fewer
940
+ epochs in the final fine-tuning stage. However, as shown in
941
+ Fig. 5 (c), too large ρ brings accuracy degradation, so ρ can
942
+ be considered as a parameter that controls the trade-off be-
943
+ tween the speed of imposing sparsity and task performance.
944
+ Visualization. Fig. 6 illustrates the sparsity patterns in
945
+ the pruned ViT models after performing our GOHSP ap-
946
+ proach. It is seen that three types of structured sparsity pat-
947
+ terns (head-level sparsity, column-level sparsity in the head
948
+ and column-level sparsity in the MLP) are imposed on the
949
+ 0
950
+ 10
951
+ 20
952
+ 30
953
+ 40
954
+ 50
955
+ 60
956
+ Epoch
957
+ 0
958
+ 250
959
+ 500
960
+ Loss
961
+ (a) Curves of training loss
962
+ =0.001
963
+ =0.002
964
+ =0.0005
965
+ 0
966
+ 10
967
+ 20
968
+ 30
969
+ 40
970
+ 50
971
+ 60
972
+ Epoch
973
+ 0
974
+ 50
975
+ 100
976
+ L2-Norm
977
+ (b) Curves of sparsity strength
978
+ =0.001
979
+ =0.002
980
+ =0.0005
981
+ 0
982
+ 10
983
+ 20
984
+ 30
985
+ 40
986
+ 50
987
+ 60
988
+ Epoch
989
+ 50
990
+ 75
991
+ 100
992
+ Top-1 (%)
993
+ (c) Curves of test accuracy
994
+ =0.001
995
+ =0.002
996
+ =0.0005
997
+ 40
998
+ 50
999
+ 60
1000
+ 0
1001
+ 3
1002
+ 6
1003
+ 40
1004
+ 50
1005
+ 60
1006
+ 96
1007
+ 97
1008
+ 98
1009
+ Figure 5: Effect of ρ on the structured pruning procedure. ρ
1010
+ controls the trade-off between the speed of imposing sparsity
1011
+ and task performance.
1012
+ pruned models. Such pruning can be more effective on hard-
1013
+ ware than the unstructured pruning methods.
1014
+ Block9
1015
+ Block10
1016
+ Block11
1017
+ Multi-Head Attention Layer
1018
+ MLP Layer
1019
+ Block9
1020
+ Block10
1021
+ Block11
1022
+ Multi-Head Attention Layer
1023
+ MLP Layer
1024
+ Block9
1025
+ Block10
1026
+ Block11
1027
+ Multi-Head Attention Layer
1028
+ MLP Layer
1029
+ Figure 6: Visualization of the imposed structured sparsity on
1030
+ the DeiT-Small model. The columns and heads with lighter
1031
+ color are pruned. Our method can prune columns (Block9,
1032
+ Block10, and Block11), and heads (Block10, Block11) of
1033
+ the Multi-Head Attention layer. On the other hand, we can
1034
+ prune columns of MLP layers in all the blocks.
1035
+ Why
1036
+ Douglas—Rachford
1037
+ splitting
1038
+ method?
1039
+ As
1040
+ shown in our Optimization section, the iterative Dou-
1041
+ glas—Rachford splitting technique is adopted to solve Eq.
1042
+ 5. Such choice is built on two reasons. 1) Convergence:
1043
+ Douglas—Rachford splitting method is a primal-dual
1044
+ optimization method that enjoys fast convergence speed.
1045
+ According to (Boyd, Parikh, and Chu 2011), within a few
1046
+ iterations it can provide satisfied solution for large-scale
1047
+ problems – particularly attractive for DNN applications.
1048
+ More specifically for this work, the fast convergence of
1049
+ Douglas—Rachford splitting method can avoid gradient
1050
+ explosion problem introduced by the additional sparsity
1051
+ loss in Eq. 9. 2) Flexibility: Douglas—Rachford splitting
1052
+ method, by its nature, divides the original difficult optimiza-
1053
+ tion problem into several less complicated sub-problems,
1054
+ each of which can be then addressed independently. This
1055
+ divide-and-conquer property is very suitable for optimizing
1056
+ the heterogeneous structured pruning of ViT, which explores
1057
+ the different types of structured sparsity across different
1058
+ attention heads and MLPs (Eq. 10 and 11).
1059
+ Conclusion
1060
+ In this paper we propose GOHSP, a unified framework to
1061
+ perform graph and optimization-based heterogeneous struc-
1062
+ tured pruning for vision transformers. By using graph-based
1063
+ ranking and leveraging the advanced optimization tech-
1064
+ nique, our approach can efficiently impose different types
1065
+ of structured sparse patterns on the vision transformers with
1066
+ high compression rate and task performance. Our experi-
1067
+ ments show that, on ImageNet, with 30 − 50% sparsity,
1068
+ GOHSP compresses the DeiT-Tiny and DeiT-Small mod-
1069
+ els with minor or no loss in accuracy and with ∼ 25 im-
1070
+ provement in rum-time efficiency. Finally, we compress ViT-
1071
+ Small up to 80% on CIFAR10 with minor loss in accuracy.
1072
+
1073
+ References
1074
+ Anwar, S.; Hwang, K.; and Sung, W. 2017. Structured prun-
1075
+ ing of deep convolutional neural networks. ACM Journal
1076
+ on Emerging Technologies in Computing Systems (JETC),
1077
+ 13(3): 1–18.
1078
+ Bakhtiarnia, A.; Zhang, Q.; and Iosifidis, A. 2021. Single-
1079
+ Layer Vision Transformers for More Accurate Early Exits
1080
+ with Less Overhead. arXiv preprint arXiv:2105.09121.
1081
+ Boyd, S.; Parikh, N.; and Chu, E. 2011. Distributed opti-
1082
+ mization and statistical learning via the alternating direc-
1083
+ tion method of multipliers. Now Publishers Inc.
1084
+ Brown, T. B.; Mann, B.; Ryder, N.; Subbiah, M.; Kaplan, J.;
1085
+ Dhariwal, P.; Neelakantan, A.; Shyam, P.; Sastry, G.; Askell,
1086
+ A.; et al. 2020.
1087
+ Language models are few-shot learners.
1088
+ arXiv preprint arXiv:2005.14165.
1089
+ Chen, T.; Cheng, Y.; Gan, Z.; Yuan, L.; Zhang, L.; and
1090
+ Wang, Z. 2021.
1091
+ Chasing sparsity in vision transformers:
1092
+ An end-to-end exploration. Advances in Neural Information
1093
+ Processing Systems, 34.
1094
+ Deng, J.; Dong, W.; Socher, R.; Li, L.-J.; Li, K.; and Fei-
1095
+ Fei, L. 2009. Imagenet: A large-scale hierarchical image
1096
+ database. In 2009 IEEE conference on computer vision and
1097
+ pattern recognition, 248–255. Ieee.
1098
+ Devlin, J.; Chang, M.-W.; Lee, K.; and Toutanova, K. 2018.
1099
+ Bert: Pre-training of deep bidirectional transformers for lan-
1100
+ guage understanding. arXiv preprint arXiv:1810.04805.
1101
+ Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn,
1102
+ D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.;
1103
+ Heigold, G.; Gelly, S.; et al. 2020. An image is worth 16x16
1104
+ words: Transformers for image recognition at scale. arXiv
1105
+ preprint arXiv:2010.11929.
1106
+ Eckstein, J.; and Bertsekas, D. P. 1992.
1107
+ On the Dou-
1108
+ glas—Rachford splitting method and the proximal point al-
1109
+ gorithm for maximal monotone operators.
1110
+ Mathematical
1111
+ Programming, 55(1): 293–318.
1112
+ Erkan, G.; and Radev, D. R. 2004. Lexrank: Graph-based
1113
+ lexical centrality as salience in text summarization. Journal
1114
+ of artificial intelligence research, 22: 457–479.
1115
+ Han, S.; Mao, H.; and Dally, W. J. 2015.
1116
+ Deep com-
1117
+ pression: Compressing deep neural networks with pruning,
1118
+ trained quantization and huffman coding.
1119
+ arXiv preprint
1120
+ arXiv:1510.00149.
1121
+ He, Y.; Liu, P.; Wang, Z.; Hu, Z.; and Yang, Y. 2019. Filter
1122
+ pruning via geometric median for deep convolutional neu-
1123
+ ral networks acceleration. In Proceedings of the IEEE/CVF
1124
+ Conference on Computer Vision and Pattern Recognition,
1125
+ 4340–4349.
1126
+ He, Y.; Zhang, X.; and Sun, J. 2017. Channel pruning for ac-
1127
+ celerating very deep neural networks. In Proceedings of the
1128
+ IEEE international conference on computer vision, 1389–
1129
+ 1397.
1130
+ Kim, Y.-D.; Park, E.; Yoo, S.; Choi, T.; Yang, L.; and Shin,
1131
+ D. 2016.
1132
+ Compression of deep convolutional neural net-
1133
+ works for fast and low power mobile applications. In Inter-
1134
+ national Conference on Learning Representations.
1135
+ Lin, M.; Ji, R.; Wang, Y.; Zhang, Y.; Zhang, B.; Tian, Y.;
1136
+ and Shao, L. 2020. Hrank: Filter pruning using high-rank
1137
+ feature map. In Proceedings of the IEEE/CVF Conference
1138
+ on Computer Vision and Pattern Recognition, 1529–1538.
1139
+ Liu, N.; Ma, X.; Xu, Z.; Wang, Y.; Tang, J.; and Ye, J.
1140
+ 2020. AutoCompress: An automatic DNN structured prun-
1141
+ ing framework for ultra-high compression rates. In Proceed-
1142
+ ings of the AAAI Conference on Artificial Intelligence, vol-
1143
+ ume 34, 4876–4883.
1144
+ Liu, Z.; Mu, H.; Zhang, X.; Guo, Z.; Yang, X.; Cheng, K.-T.;
1145
+ and Sun, J. 2019. Metapruning: Meta learning for automatic
1146
+ neural network channel pruning. In Proceedings of the IEEE
1147
+ International Conference on Computer Vision, 3296–3305.
1148
+ Liu, Z.; Sun, M.; Zhou, T.; Huang, G.; and Darrell, T. 2018.
1149
+ Rethinking the value of network pruning.
1150
+ arXiv preprint
1151
+ arXiv:1810.05270.
1152
+ Lou, Q.; Hsu, Y.-C.; Uzkent, B.; Hua, T.; Shen, Y.; and Jin,
1153
+ H. 2022. Lite-MDETR: A Lightweight Multi-Modal Detec-
1154
+ tor. In Proceedings of the IEEE/CVF Conference on Com-
1155
+ puter Vision and Pattern Recognition, 12206–12215.
1156
+ Meng, L.; Li, H.; Chen, B.-C.; Lan, S.; Wu, Z.; Jiang, Y.-
1157
+ G.; and Lim, S.-N. 2022. AdaViT: Adaptive Vision Trans-
1158
+ formers for Efficient Image Recognition. In Proceedings of
1159
+ the IEEE/CVF Conference on Computer Vision and Pattern
1160
+ Recognition, 12309–12318.
1161
+ Mihalcea, R.; and Tarau, P. 2004. Textrank: Bringing order
1162
+ into text. In Proceedings of the 2004 Conference on Empir-
1163
+ ical Methods in Natural Language Processing, 404–411.
1164
+ Mocanu, D. C.; Mocanu, E.; Stone, P.; Nguyen, P. H.;
1165
+ Gibescu, M.; and Liotta, A. 2018. Scalable training of ar-
1166
+ tificial neural networks with adaptive sparse connectivity in-
1167
+ spired by network science. Nature Communications, 9(1):
1168
+ 1–12.
1169
+ Page, L.; Brin, S.; Motwani, R.; and Winograd, T. 1999. The
1170
+ PageRank citation ranking: Bringing order to the web. Tech-
1171
+ nical report, Stanford InfoLab.
1172
+ Pan, Y.; Xu, J.; Wang, M.; Ye, J.; Wang, F.; Bai, K.; and Xu,
1173
+ Z. 2019. Compressing recurrent neural networks with tensor
1174
+ ring for action recognition. In Proceedings of the AAAI Con-
1175
+ ference on Artificial Intelligence, volume 33, 4683–4690.
1176
+ Rao, Y.; Zhao, W.; Liu, B.; Lu, J.; Zhou, J.; and Hsieh,
1177
+ C.-J. 2021.
1178
+ DynamicViT: Efficient Vision Transform-
1179
+ ers with Dynamic Token Sparsification.
1180
+ arXiv preprint
1181
+ arXiv:2106.02034.
1182
+ Seneta, E. 2006. Non-negative matrices and Markov chains.
1183
+ Springer Science & Business Media.
1184
+ Tiwari, R.; Bamba, U.; Chavan, A.; and Gupta, D. 2021.
1185
+ ChipNet: Budget-Aware Pruning with Heaviside Continuous
1186
+ Approximations. In International Conference on Learning
1187
+ Representations.
1188
+ Touvron, H.; Cord, M.; Douze, M.; Massa, F.; Sablayrolles,
1189
+ A.; and J´egou, H. 2021. Training data-efficient image trans-
1190
+ formers & distillation through attention.
1191
+ In International
1192
+ Conference on Machine Learning, 10347–10357. PMLR.
1193
+ Uzkent, B.; and Ermon, S. 2020. Learning when and where
1194
+ to zoom with deep reinforcement learning. In Proceedings
1195
+
1196
+ of the IEEE/CVF conference on computer vision and pattern
1197
+ recognition, 12345–12354.
1198
+ Uzkent, B.; Yeh, C.; and Ermon, S. 2020. Efficient object
1199
+ detection in large images using deep reinforcement learn-
1200
+ ing. In Proceedings of the IEEE/CVF winter conference on
1201
+ applications of computer vision, 1824–1833.
1202
+ Wang, Y.; Huang, R.; Song, S.; Huang, Z.; and Huang, G.
1203
+ 2021. Not All Images are Worth 16x16 Words: Dynamic
1204
+ Vision Transformers with Adaptive Sequence Length. arXiv
1205
+ preprint arXiv:2105.15075.
1206
+ Wen, W.; Wu, C.; Wang, Y.; Chen, Y.; and Li, H. 2016.
1207
+ Learning structured sparsity in deep neural networks. Ad-
1208
+ vances in neural information processing systems, 29: 2074–
1209
+ 2082.
1210
+ Xu, Y.; Zhang, Z.; Zhang, M.; Sheng, K.; Li, K.; Dong, W.;
1211
+ Zhang, L.; Xu, C.; and Sun, X. 2022. Evo-vit: Slow-fast
1212
+ token evolution for dynamic vision transformer. In Proceed-
1213
+ ings of the AAAI Conference on Artificial Intelligence, vol-
1214
+ ume 36, 2964–2972.
1215
+ Ye, J.; Lu, X.; Lin, Z.; and Wang, J. Z. 2018. Rethinking the
1216
+ smaller-norm-less-informative assumption in channel prun-
1217
+ ing of convolution layers. arXiv preprint arXiv:1802.00124.
1218
+ Yu, R.; Li, A.; Chen, C.-F.; Lai, J.-H.; Morariu, V. I.; Han,
1219
+ X.; Gao, M.; Lin, C.-Y.; and Davis, L. S. 2018. Nisp: Prun-
1220
+ ing networks using neuron importance score propagation. In
1221
+ Proceedings of the IEEE/CVF Conference on Computer Vi-
1222
+ sion and Pattern Recognition, 9194–9203.
1223
+ Yu, S.; Chen, T.; Shen, J.; Yuan, H.; Tan, J.; Yang, S.; Liu, J.;
1224
+ and Wang, Z. 2022. Unified Visual Transformer Compres-
1225
+ sion. In International Conference on Learning Representa-
1226
+ tions.
1227
+ Yu, X.; Liu, T.; Wang, X.; and Tao, D. 2017. On compress-
1228
+ ing deep models by low rank and sparse decomposition. In
1229
+ Proceedings of the IEEE Conference on Computer Vision
1230
+ and Pattern Recognition, 7370–7379.
1231
+ Zhai, X.; Kolesnikov, A.; Houlsby, N.; and Beyer, L.
1232
+ 2021.
1233
+ Scaling vision transformers.
1234
+ arXiv preprint
1235
+ arXiv:2106.04560.
1236
+ Zhou, D.; Kang, B.; Jin, X.; Yang, L.; Lian, X.; Jiang, Z.;
1237
+ Hou, Q.; and Feng, J. 2021. Deepvit: Towards deeper vision
1238
+ transformer. arXiv preprint arXiv:2103.11886.
1239
+ Zhu, M.; and Gupta, S. 2017. To prune, or not to prune: ex-
1240
+ ploring the efficacy of pruning for model compression. arXiv
1241
+ preprint arXiv:1710.01878.
1242
+ Zhuang, Z.; Tan, M.; Zhuang, B.; Liu, J.; Guo, Y.; Wu, Q.;
1243
+ Huang, J.; and Zhu, J. 2018. Discrimination-aware channel
1244
+ pruning for deep neural networks. In Advances in Neural
1245
+ Information Processing Systems, 875–886.
1246
+
LNE4T4oBgHgl3EQf8A7c/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
LtFRT4oBgHgl3EQf2Di0/content/tmp_files/2301.13659v1.pdf.txt ADDED
@@ -0,0 +1,1263 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 1
2
+ Spyker: High-performance Library for Spiking
3
+ Deep Neural Networks
4
+ Shahriar Rezghi Shirsavar†‡, Mohammad-Reza A. Dehaqani†‡,
5
+ †School of Electrical and Computer Engineering, College of Engineering, University of Tehran, Tehran, Iran
6
+ {shahriar.rezghi, dehaqani}@ut.ac.ir
7
+ ‡School of Cognitive Sciences, Institute for Research in Fundamental Sciences (IPM), Tehran, Iran
8
+ ∗Corresponding author: Mohammad-Reza A. Dehaqani, [email protected]
9
+ Abstract—Spiking neural networks (SNNs) have been recently
10
+ brought to light due to their promising capabilities. SNNs
11
+ simulate the brain with higher biological plausibility compared
12
+ to previous generations of neural networks. Learning with fewer
13
+ samples and consuming less power are among the key features
14
+ of these networks. However, the theoretical advantages of SNNs
15
+ have not been seen in practice due to the slowness of simulation
16
+ tools and the impracticality of the proposed network structures.
17
+ In this work, we implement a high-performance library named
18
+ Spyker using C++/CUDA from scratch that outperforms its
19
+ predecessor. Several SNNs are implemented in this work with
20
+ different learning rules (spike-timing-dependent plasticity and
21
+ reinforcement learning) using Spyker that achieve significantly
22
+ better runtimes, to prove the practicality of the library in the
23
+ simulation of large-scale networks. To our knowledge, no such
24
+ tools have been developed to simulate large-scale spiking neural
25
+ networks with high performance using a modular structure.
26
+ Furthermore, a comparison of the represented stimuli extracted
27
+ from Spyker to recorded electrophysiology data is performed
28
+ to demonstrate the applicability of SNNs in describing the
29
+ underlying neural mechanisms of the brain functions. The aim
30
+ of this library is to take a significant step toward uncovering the
31
+ true potential of the brain computations using SNNs.
32
+ Index
33
+ Terms—Spiking
34
+ Neural
35
+ Network,
36
+ Learning
37
+ Rules,
38
+ C++/CUDA, Modular Structure, Biological Plausibility
39
+ I. INTRODUCTION
40
+ The human brain can operate with amazing robustness and
41
+ energy efficiency. Artificial neural networks (ANNs) aim at
42
+ modeling the brain, and three generations of these networks
43
+ have been developed. Each generation of ANNs improves the
44
+ quality of the modeling of the brain compared to the last. The
45
+ first generation of ANNs makes use of the McCulloch-Pitts
46
+ neurons [1]. Although these neurons are inspired by biological
47
+ neurons, time dynamics are not considered in this model, and
48
+ the learning rules proposed for them lack power and biological
49
+ plausibility. These neurons were used in multi-layer perceptron
50
+ (MLPs) [2] and Hopfield [3] networks.
51
+ The second generation of ANNs uses a continuous activa-
52
+ tion function (ReLU [4] and sigmoid [5], for example) instead
53
+ of thresholding, which makes them suitable for processing
54
+ analog signals. They have attracted the attention of researchers
55
+ in recent years and were able to reach high accuracies [6], [7]
56
+ (even surpassing humans) and win different challenges [8].
57
+ Despite the success of DNNs, there are structural differences
58
+ between these networks and the human brain. Lack of temporal
59
+ dynamics, using analog signals for network propagation and
60
+ activation functions, learning rules without biological roots,
61
+ and the need for large amounts of data [9] and energy [10] to
62
+ achieve acceptable results are among these differences.
63
+ The third generation of neural networks is spiking neural
64
+ networks (SNNs). The neural models used in these networks
65
+ simulate biological neurons more accurately, and the coding
66
+ mechanisms used in these networks are found in neural
67
+ communications. Furthermore, the learning rules used in these
68
+ networks have been discovered in the brain [11]–[13]. Having
69
+ lower energy consumption, learning with fewer samples, and
70
+ solving more complicated tasks due to time dynamics (several
71
+ electrophysiological studies emphasize the role of temporal
72
+ dynamics in neural coding [14], [15]) are some of the advan-
73
+ tages of SNNs compared to the second generation of ANNs.
74
+ SNNs can be used to solve machine learning tasks, study and
75
+ explore brain functionality, and run on specialized hardware
76
+ with low power consumption. The research being done on
77
+ these networks aims to address the disadvantages of DNNs
78
+ with more realistic modeling of the brain functionality.
79
+ Several high-performance well-established frameworks like
80
+ PyTorch [16], TensorFlow [17], and MXNet [18] have been
81
+ developed for DNNs in recent years. These libraries have en-
82
+ abled DNNs to achieve new highs in solving machine learning
83
+ tasks. SNNs are not yet comparable to DNNs due to the lack
84
+ of fast simulation tools. There have been some attempts, like
85
+ SpykeTorch [19] and BindsNet [20]. SpykeTorch, written on
86
+ top of the PyTorch framework, is a simulator for large-scale
87
+ spiking neural networks (SDNNs). However, it has a slow
88
+ runtime, and training even simple networks can take up to days
89
+ to complete. To our knowledge, Spyker is the first toolbox to
90
+ simulate large-scale networks with high performance, is easy
91
+ to use, has the flexibility to be used in multiple languages, and
92
+ has the compatibility to integrate with other commonly used
93
+ tools. In order to fill this need, we have developed Spyker.
94
+ Spyker is a C++/CUDA library written from scratch with both
95
+ C++ and Python interfaces and support for dense and sparse
96
+ structures. Although Spyker is a stand-alone library, it has a
97
+ highly flexible API and can work with PyTorch tensors and
98
+ Numpy arrays. Figure 1 shows an overview of the library. In
99
+ order to increase performance, small-sized integers are used
100
+ alongside floating-point numbers. It also uses highly-optimized
101
+ low-level back-end libraries such as OneDNN and cuDNN to
102
+ speed up heavy computations such as convolutions and matrix
103
+ multiplications. Spyker can be compiled on various CPUs to be
104
+ arXiv:2301.13659v1 [cs.CV] 31 Jan 2023
105
+
106
+ 2
107
+ optimized locally and take advantage of native CPU-specific
108
+ instructions.
109
+ Spiking neural networks are made of different building
110
+ blocks (see [21] for more details). The first block is the
111
+ modeling of the biological neurons. Some examples of this
112
+ are leaky integrate-and-fire [22], spike-response model [23],
113
+ and Izhikevich model [24]. Another building block is neural
114
+ coding, which can be rate coding [25], temporal coding, phase
115
+ coding and synchrony coding [26], or other coding schemes.
116
+ The final building block is the learning mechanism. Examples
117
+ of these mechanisms are STDP [27], [28], R-STDP [29],
118
+ backpropagation [30], and conversion from ANNs to SNNs
119
+ [31]. Spyker has a modular implementation of these three
120
+ blocks that enables its users to build SNNs.
121
+ Spyker provides SNN functionality with a high-performance
122
+ and easy-to-use interface with an open-source and permissive
123
+ license. It can run on CPU and CUDA devices and has
124
+ both dense and sparse interfaces. The library introduces new
125
+ features and fixes most of the shortcomings of its prede-
126
+ cessor. The improvements include adding batch processing,
127
+ strided convolutions, internal padding for convolutions, fully
128
+ connected layers, and the rate coding mechanism. Compared
129
+ to its predecessor, the interface of the library is simpler,
130
+ closer to the current API of deep learning libraries, and more
131
+ straightforward to use. In this work, several successful network
132
+ structures are implemented using this library to prove its
133
+ operability, its runtime speed is compared to SpykeTorch, and
134
+ the results indicate Spyker can run up to eight times faster.
135
+ The proposed work is able to reduce the gap between SNNs
136
+ and DNNs and bring us a step closer to uncovering the true
137
+ potential of spiking neural networks.
138
+ We start with a description of dimensionality of the input
139
+ arrays and how the spike trains are implemented in the library.
140
+ Afterward, we provide an explanation of different building
141
+ blocks of SNNs and how they are implemented in Spyker and
142
+ modeled in the interface. Then, we implement network struc-
143
+ tures that have been succesful to prove its operatibility, and we
144
+ compare the performance of the library to its predecessor on
145
+ these networks. Furthermore, comparison of the represented
146
+ stimuli extracted from Spyker to recorded electrophisiology
147
+ data is performed to demonstrate the applicability of SNNs
148
+ in describing the underlying neural mechanisms of the brain
149
+ functions. Finally, we demonstrate an example usage of the
150
+ library and discuss the impacts of this work and how it can
151
+ be further improved.
152
+ II. METHODS
153
+ The interface of the Spyker can be better explained when the
154
+ classes and methods of the interface are grouped by building
155
+ blocks of SNNs. The categories are feature enhancement,
156
+ neural coding, neural model, and learning. In this section, the
157
+ structure of the input to the network is explained. Afterward,
158
+ the sparse and the dense interfaces are compared. Finally, the
159
+ building blocks of the library are discussed in detail.
160
+ A. Network Input
161
+ Arrays passed through convolutional neural networks that
162
+ process images are often four-dimensional arrays composed
163
+ of batch size (B or N), number of channels (C), image height
164
+ (H), and image width (W). The order can either be BCHW
165
+ or BHWC (or NCHW or NHWC). SNNs have temporal
166
+ dynamics, and it is implemented as a dimension that represents
167
+ time steps in Spyker. The library implements five-dimensional
168
+ arrays with BTCHW order (T being the time steps). Since
169
+ DNNs process analog signals, data types used in these net-
170
+ works are (usually four-byte) floating-point numbers. This data
171
+ type can be computationally expensive compared to a small-
172
+ sized integer type and take up more space in the memory.
173
+ Since SNNs process binary signals, Spyker can optionally use
174
+ eight-bit (or wider) integers alongside floating-point numbers
175
+ to improve performance further.
176
+ B. Dense vs Sparse interface
177
+ The dense interface of Spyker uses the fully allocated
178
+ memory buffers that are used in neural network computations.
179
+ However, the sparse interface only needs to hold the indices
180
+ of the spikes. Conversion between dense and sparse interfaces
181
+ are provided in the library. The sparse interface has some
182
+ advantages compared to the dense interface. In the dense
183
+ interface, the time consumed by each operation is a function
184
+ of the size of each of the 5 dimensions. However, in the sparse
185
+ interface, it depends on the number of spikes. This means both
186
+ memory and time consumed will be greatly reduced when
187
+ processing sparser signals. Furthermore, since neurons fire at
188
+ most once when using rank order coding, the increment of the
189
+ number of time steps will have a smaller effect in the sparse
190
+ interface compared to the dense interface.
191
+ C. Feature Enhancement
192
+ A transformation can be used to enhance features of the in-
193
+ put signal (image) before the neural coding process [32]–[34].
194
+ This results in highlighted features having higher intensities
195
+ and appearing in earlier time steps, meaning more excitation.
196
+ Feature enhancement is done through filtering the input here.
197
+ Various filters are supported in Spyker, and they are introduced
198
+ in the following subsections.
199
+ 1) Difference of Gaussian Filter: The first filter is the Dif-
200
+ ference of Gaussian (DoG). This filter increases the intensities
201
+ of edges and other details in the image (see Figure 2 for an
202
+ example) [35]. It approximates the center-surround properties
203
+ of the ganglion cells of the retina [36] (see also [37], [38]).
204
+ This operation is implemented as spyker.DoG(size, filters, pad
205
+ , device) where size is the size of the width and the height
206
+ of the filter, filters is a list of DoG filter descriptions (each
207
+ description takes in two standard deviations), pad is the size
208
+ of the padding of the image, and device is the device the filter
209
+ will run on (CPU, GPU or others).
210
+ 2) Gabor Filter: The following filter is the Gabor filter
211
+ that determines the presence of specific frequency in content
212
+ in a specific direction in the image. Research Indicates [39]
213
+ that the Gabor filter is used in the human visual cortex. The
214
+ Gabor filter is implemented as spyker.Gabor(size, filters, pad
215
+ , device). The parameters of this class are the same as the
216
+ DoG class, but the filters are Gabor filter descriptions, and
217
+ each description takes in sigma, theta, gamma, lambda, and
218
+ psi.
219
+
220
+ 3
221
+ Numpy Array
222
+ PyTorch Tensor
223
+ Numpy Array
224
+ PyTorch Tensor
225
+ Feature Enhancement
226
+ Neural Coding
227
+ Neural Model
228
+ Learning
229
+ T=0
230
+ T=1
231
+ T=2
232
+ T=3
233
+ A+
234
+ A-
235
+ Fig. 1: Overview of the Spyker library. Spyker API supports PyTorch tensors and Numpy arrays as well as a built-in data
236
+ wrapper. The output of Spyker operations have the same container type as the input. The functionality of Spyker can be grouped
237
+ into subcategories shown in the figure.
238
+ 3) Laplacian of Gaussian Filter: The Laplacian of Gaus-
239
+ sian (LoG) layer is also implemented in Spyker, and it is ap-
240
+ proximated using two DoG filters. An LoG filter with standard
241
+ deviation σ can be approximated using two DoG filters with
242
+
243
+
244
+ 2, σ/
245
+
246
+ 2) and (σ/
247
+
248
+ 2, σ
249
+
250
+ 2) standard deviations. This
251
+ filter exists in Spyker as spyker.LoG(size, stds, pad, device)
252
+ where stds are a list of standard deviations needed to describe
253
+ multiple LoG filters.
254
+ 4) Shape of the Filters: The previously explained filters
255
+ have kernel size Kc × Kh × Kw, which are square kernels
256
+ (Kh = Kw). The input can have B × Ci × Hi × Wi shape
257
+ which corresponds to batch, channels, height, and width of the
258
+ input, respectively. The output will have B × Co × Ho × Wo
259
+ shape where:
260
+ Co = Ci × Kc
261
+ Ho = Hi + 2 × Ph − Kh + 1
262
+ Wo = Wi + 2 × Ph − Kw + 1
263
+ (1)
264
+ and Ph and Pw are height and width padding of the filter. The
265
+ Kc filters are applied to each channel separately.
266
+ 5) Zero-phase Component Analysis:
267
+ Final implemented
268
+ layer is zero-phase component analysis (ZCA) Whitening.
269
+ It has been suggested [34] that this transformation can im-
270
+ prove the accuracy of SNNs on real-world images. Spyker
271
+ implements an efficient version of ZCA whitening by taking
272
+ advantage of routines from highly optimized linear algebra
273
+ libraries (BLAS and LAPACK) that operate on symmetric
274
+ matrices. This layer is implemented as spyker.ZCA class
275
+ which has a fit(array, epsilon) and a call function.
276
+ D. Neural Coding
277
+ SNNs process spike trains, but the input consists of analog
278
+ values (for example, images are made of pixel values). In order
279
+ to make these inputs suitable for the network, a conversion
280
+ scheme is needed. The mapping from stimuli to neural re-
281
+ sponses is called neural coding [40]. Coding schemes imple-
282
+ mented in Spyker are explained in the following subsections.
283
+ 1) Rate Coding: Out of several coding schemes suggested,
284
+ rate coding is widely used where the rate of firing of the
285
+ neurons represents information. In this scheme, the rate of
286
+ firing is dependent on the intensity of the input value (higher
287
+ intensity corresponds to faster firing) [25]. The exact time
288
+ of firing in each neuron is stochastic in nature and may be
289
+ modeled with a Poisson distribution. A lengthy window of
290
+ time is required to transmit the information in this coding,
291
+ and the spikes are not quite sparse.
292
+ 2) Temporal Coding: Another popular coding scheme is
293
+ temporal coding [41]. Recordings in the primary visual cortex
294
+ show [42] that the response latency decreases with the stimulus
295
+ contrast. This coding scheme can convey information through
296
+ the timings of the spikes. Multiple forms of this scheme have
297
+ been proposed, including rank order coding [43]. Instead of
298
+ computing the exact timing of each spike, the timings are
299
+ computed relative to one another in rank order coding. This
300
+ relative (instead of exact) timing can increase invariance to
301
+ changes in the input intensity and contrast [43]. It has been
302
+ suggested [44] that temporal coding might be more efficient
303
+ in some situations.
304
+ 3) Coding in Spyker: Spyker supports rank order and rate
305
+ coding. The concept of time is implemented with spikes
306
+ occuring in time steps in this library. Rank order coding maps
307
+ higher intensities to earlier time steps of a neuron firing. In
308
+ order to calculate the time step the neuron will fire in, Spyker
309
+ sorts the intensity values by default. This calculates rank order
310
+ between spikes, and the spikes will be distributed among
311
+ time steps evenly. The sorting operation is computationally
312
+
313
+ S
314
+ P
315
+ Y
316
+ K
317
+ E
318
+ R4
319
+ T=0 T=1 T=2 T=3
320
+ B&W Image
321
+ B&W Image
322
+ DoG Filtered
323
+ Gabor Filtered
324
+ T=0
325
+ T=0
326
+ T=1
327
+ T=1
328
+ T=2
329
+ T=2
330
+ T=3
331
+ T=3
332
+ Input Image
333
+ (Gray or HSV)
334
+ Feature
335
+ Enhancement
336
+ Encoded input data ready to be processed by the network
337
+ Neural Coding
338
+ Fig. 2: The figure shows a black and white image being filtered by DoG and Gabor filters. The theta parameter of the Gabor
339
+ filter is set to -15 degrees. Then the images are coded using rank order coding into four time steps. Spikes are shown with
340
+ white color on a black background through time steps. Spikes carry on from the previous to the current time step (cumulative
341
+ structure).
342
+ expensive (specially on GPUs), and optionally, it can be
343
+ disabled to have runtime improvements (however, accuracy
344
+ might be affected). Since processing time steps sequantially is
345
+ inefficient and time-consuming, Spyker processes all the time
346
+ steps at once. To this end, when a neuron fires in time step ti,
347
+ it will also fire at time steps ti+1, ti+2, ..., tn where n is the
348
+ number of time steps. An example of this cumulative structure
349
+ can be seen in Figure 2.
350
+ E. Neural Model
351
+ Once the input is filtered and coded, it gets processed
352
+ by the network. The network is built using fully connected,
353
+ convolution, integrate-and-fire (IF) activation, pooling, and
354
+ padding layers. These operations are explained in the follow-
355
+ ing subsections.
356
+ 1) Convolution: The integrate-and-fire mechanism is im-
357
+ plemented by combining convolution and the IF activation
358
+ layer. The internal potentials of the neurons are computed
359
+ using convolution operation, and the IF activation operation
360
+ produces spikes where neurons have a potential higher than
361
+ a specified threshold. Multiple layers can be assembled and
362
+ stacked on top of one another to create deeper structures.
363
+ The convolution layer has a kernel with Co×Ci×Kh×Kw
364
+ shape. the synaptic weights are initialized randomly with
365
+ a normal distribution. It performs two-dimensional convo-
366
+ lution with support for padding and stride. The input has
367
+ B ×T ×Ci ×Hi ×Wi shape which corresponds to batch, time
368
+ steps, channels, height, and width of the input, respectively.
369
+ The output has B × T × Co × Ho × Wo shape where:
370
+ Ho = ⌊Hi + 2 × Ph − Kh
371
+ Sh
372
+ ⌋ + 1
373
+ Wo = ⌊Wi + 2 × Pw − Kw
374
+ Sw
375
+ ⌋ + 1
376
+ (2)
377
+ And Ph, Pw, Sh, Sw are the height and width of convolution
378
+ padding and stride. Padding increases the size of the two-
379
+ dimensional input before convolution operation by expanding
380
+ the edges of the input and filling in the new space with a
381
+ constant value (usually zero). Stride is the number of steps
382
+ the convolution window takes when it moves on the image.
383
+ The output of the convolution layers are internal potentials
384
+ of neurons that need to be passed through an IF activation
385
+ layer to become output spike trains. This layer is imple-
386
+ mented with spyker.Conv(insize, outsize, kernel, stride, pad,
387
+ mean, std, device) class in Spyker.
388
+ 2) Fully Connected: The fully connected layer is combined
389
+ with the IF activation to model the IF neurons, much similar
390
+ to what happens in the convolution layers. This layer has a
391
+ kernel with I × O shape. The synaptic weights are initialized
392
+
393
+ 5
394
+ randomly with a normal distribution. The input has B ×T ×I
395
+ which corresponds to batch, time steps, and input size, respec-
396
+ tively. The output has B × T × O shape. The fully connected
397
+ layer is represeneted by spyker.FC(insize, outsize, mean, std,
398
+ device) in the library.
399
+ 3) Pooling: The pooling layer performs two-dimensional
400
+ max pooling operation with a window size ofLh×Lw, a stride
401
+ of Sh ×Sw, and a padding of Ph, Pw. The input has B ×T ×
402
+ Ci×Hi×Wi shape and the output has B×T ×Co×Ho×Wo
403
+ shape where:
404
+ Ho = ⌊Hi + 2 × Ph − Lh
405
+ Sh
406
+ ⌋ + 1
407
+ Wo = ⌊Wi + 2 × Pw − Lw
408
+ Sw
409
+ ⌋ + 1
410
+ (3)
411
+ The interface of Spyker has the spyker.pool(array, kernel,
412
+ stride, pad, rates) function to run the pooling operation on the
413
+ input given the kernel, stride, and padding size. rates argument
414
+ is the rate of firing of the neurons when rate coding is used.
415
+ The pooling operation selects neurons that fire earlier when
416
+ rank order coding is used, and selects neurons that have a
417
+ higher firing rate when rate coding is used.
418
+ F. Learning
419
+ Learning in the brain happens when the strength of connec-
420
+ tions change between its neurons, and this change in strength
421
+ is named synaptic plasticity [45]. Learning methods that utilize
422
+ synaptic plasticity have been developed for SNNs [27]–[29].
423
+ 1) Spike-timing-dependent Plasticity:
424
+ One widely rec-
425
+ ognized synaptic plasticity learning rule is spike-timing-
426
+ dependent plasticity (STDP) [27], [28]. STDP learning rule op-
427
+ erates by adjusting synaptic weights and utilizing the timing of
428
+ the spikes. A pre-synaptic neuron firing before (after) the post-
429
+ synaptic neuron results in a strengthed (weakened) connection.
430
+ STDP allows the neurons to extract and learn frequent features
431
+ in the input [46]. STDP layer changes synaptic weights with
432
+ stabilization:
433
+ ∆Wi,j =
434
+
435
+ A+
436
+ k (Wi,j − Lk)(Uk − Wi,j),
437
+ tj ≤ ti
438
+ A−
439
+ k (Wi,j − Lk)(Uk − Wi,j),
440
+ tj ≥ ti
441
+ (4)
442
+ where A+
443
+ k , A−
444
+ k , Lk, Uk are the positive learning rate, negative
445
+ learning rate, lower bound, and upper bound of the kth
446
+ configuration, respectively. If stabilization is not set, then the
447
+ formula becomes:
448
+ ∆Wi,j =
449
+
450
+ A+
451
+ k ,
452
+ tj ≤ ti
453
+ A−
454
+ k ,
455
+ tj ≥ ti
456
+ (5)
457
+ then the weights are computed:
458
+ W +
459
+ i,j = max(Lk, min(Uk, ∆Wi,j))
460
+ (6)
461
+ Input, neurons selected by the winner-take-all mechanism
462
+ (WTA), and the output are passed to a function belonging to
463
+ fully connected or convolution layers, and the STDP learn-
464
+ ing rule is applied. Convolution or fully connected layers
465
+ in Spyker can have multiple STDP configurations (differ-
466
+ ent learning rules, weight clipping, enabling/disabling stabi-
467
+ lizer) implemented as spyker.STDPConfig(positive, negative,
468
+ stabilize, lower, upper). Each winner neuron can be mapped
469
+ to an STDP configuration, and that neuron will be updated
470
+ using the learning rates and such that belongs to the selected
471
+ configuration. SpykeTorch creates an STDP object for each
472
+ configuration, and mapping winner neurons to different con-
473
+ figurations is done by the user. Compared to SpykeTorch,
474
+ Spyker provides a more flexible and easy to use API for
475
+ weight updating and enables batch updating, which improves
476
+ performance. Samples are processed in mini-batches which
477
+ increases performance drastically (see the results section),
478
+ and the batch update rule does not differ from single-sample
479
+ processing.
480
+ 2) Reward-modulated STDP: Another approach is using
481
+ the reinforcement (RL) learning rule. One method based on RL
482
+ is reward-modulated STDP [29]. R-STDP adjusts the STDP
483
+ such that neurons that respond correctly are rewarded, and
484
+ punished otherwise. It has been suggested [33] that when
485
+ the input has non-diagnostic frequent features that are less
486
+ effective in decision-making, R-STDP is able to discard these
487
+ features and improve the decision-making process. Since con-
488
+ volution and fully connected layers accept STDP configura-
489
+ tions as input, R-STDP can be implemented by passing two
490
+ configurations to a layer (one for rewarding and one for pun-
491
+ ishing), and mapping each winner neuron to a configuration
492
+ based on data labels. If one formulates this, ∆Wi,j will be:
493
+
494
+
495
+
496
+
497
+
498
+
499
+
500
+
501
+
502
+
503
+
504
+
505
+
506
+
507
+
508
+
509
+ A+
510
+ r (Wi,j − Lr)(Ur − Wi,j),
511
+ tpre < tpost
512
+ A−
513
+ r (Wi,j − Lr)(Ur − Wi,j),
514
+ tpre ≥ tpost
515
+ ,
516
+ if reward
517
+
518
+ A−
519
+ p (Wi,j − Lp)(Up − Wi,j),
520
+ tpre < tpost
521
+ A+
522
+ p (Wi,j − Lp)(Up − Wi,j),
523
+ tpre ≥ tpost
524
+ ,
525
+ if punish
526
+ (7)
527
+ 3) Winner-take-all and Lateral Inhibition: When a neu-
528
+ ron fires at a specific location, lateral inhibition [47], [48]
529
+ operation inhibits other neurons belonging to other neural
530
+ maps from firing in that location. Lateral inhibition for the
531
+ convolution operation can be used with spyker.inhibit(array
532
+ , thershold, inplace) functions. Winner neurons that STDP
533
+ weight updating will be performed on are selected by the
534
+ winner-take-all [49], [50] operation. WTA selects neurons that
535
+ fire earlier, and if the firing time of neurons is the same, then
536
+ the one that has a higher internal potential will be selected.
537
+ This operation is implemented with spyker.fcwta(array, radius
538
+ , count, threshold) for fully connected and spyker.convwta(
539
+ array, radius, count, threshold) for convolution operations.
540
+ III. RESULTS
541
+ In this section, we will test the performance of the library
542
+ against the SpykeTorch library. Afterward, a comparison of the
543
+ represented stimuli extracted from Spyker to recorded electro-
544
+ physiology data is conducted to demonstrate the applicability
545
+ of SNNs in describing the underlying neural mechanisms of
546
+ brain functions.
547
+ A. Library Performance
548
+ In this section, we compare the performance of the library to
549
+ SpykeTorch on two networks that classify the MNIST dataset.
550
+
551
+ 6
552
+ SpykeTorch
553
+ Spyker Python
554
+ Spyker Python Alt
555
+ Spyker C++
556
+ Spyker C++ Alt
557
+ 95
558
+ 96
559
+ 97
560
+ 98
561
+ 99
562
+ 100
563
+ Accuracy (%)
564
+ 96.72
565
+ 97.55
566
+ 97.632
567
+ 97.502
568
+ 97.606
569
+ Accuracy results for the MNIST dataset
570
+ SpykeTorch
571
+ Spyker Python
572
+ Spyker Python Alt
573
+ Spyker C++
574
+ Spyker C++ Alt
575
+ 0h
576
+ 3h
577
+ 6h
578
+ 9h
579
+ 12h
580
+ 15h
581
+ 18h
582
+ 21h
583
+ Time (s)
584
+ 21h 17m
585
+ 4h 21m
586
+ 3h 21m
587
+ 4h 7m
588
+ 3h 8m
589
+ Runtime results for the MNIST dataset
590
+ Fig. 3: Comparison plots of the runtime and accuracy of
591
+ Spyker aganist SpykeTorch on the Mozafari et al. network.
592
+ The plot on the left shows the runtime comparison of Spyker
593
+ and SpykeTorch implementations. The plot on the right also
594
+ compares accuracy of the two implementations. Comparisons
595
+ are between SpykeTorch (ST), implementation using Spyker
596
+ in Python (SP Py), alternative version using Spyker in Python
597
+ (SPA Py), and their C++ counterparts (SP C++, SPA C++). The
598
+ error bars are minimum and maximum values of the samples.
599
+ 1) R-STDP Network: The first netwrok is the Mozfari et al.
600
+ network [33] which has three convolutional layers. The first
601
+ layer is trained two times with STDP, the second layer four
602
+ times with STDP, and the third layer 680 times with R-STDP
603
+ on the training set while compuing the test accuracy at each
604
+ iteration while training the third layer. We made a small change
605
+ to the structure of the network (named Alt for alternative).
606
+ We removed the input padding from the last convolution layer
607
+ and changed its window size to 4 and the output channels
608
+ to 400. Results can be seen in Figure 3 and Table I. All the
609
+ tests are performed on Inte Core i7-9700k with 64G memory
610
+ and Nvidia Geforce GTX 1080 Ti with 12G memory (Ubuntu
611
+ 18.04).
612
+ In order to compare the results, we test whether the two-
613
+ sample mean difference confidence interval (99.9%) contains
614
+ zero. The null hypothesis is having the same means, and the
615
+ alternative is having different means. The test results indicate
616
+ that the Spyker Python implementation is faster compared
617
+ to the SpykeTorch implementation (Confidence intervals are
618
+ TABLE I: Comparisons of the the runtime and accuracy of
619
+ Spyker aganist SpykeTorch on the Mozafari et al. network.
620
+ Implementation
621
+ Time
622
+ Time
623
+ (S±SD)
624
+ Accuracy
625
+ (%±SD)
626
+ Runs
627
+ SpykeTorch
628
+ 21h17m
629
+ 76,672±916
630
+ 96.720±0.163
631
+ 12
632
+ Spyker Python
633
+ 04h49m
634
+ 15668±52
635
+ 97.550±0.169
636
+ 30
637
+ Spyker Python Alt
638
+ 03h31m
639
+ 12,114±14
640
+ 97.632±0.112
641
+ 30
642
+ Spyker C++
643
+ 03h52m
644
+ 14,869±50
645
+ 97.502±0.157
646
+ 30
647
+ [15477, 15859] and [72607, 80737] for Spyker and Spyke-
648
+ Torch respectively, showing no intersection). Furthermore,
649
+ the alternative implementation is faster both in the Python
650
+ implementation with [-3738, -3370] interval and the C++
651
+ implementation with [-3828, -3339] interval. As expected, the
652
+ C++ interface is faster compared to the Python interface with [-
653
+ 1078, -520] interval. The results for the accuracy comparisons
654
+ show that there are no significant differences ([96.932, 98.169]
655
+ and [95.996, 97.444] for Python vs SpykeTorch implementa-
656
+ tions respectively, showing intersection, [-0.89, 0.793] for C++
657
+ vs Python, [-0.649, 0.813] for Python alternative vs Python,
658
+ and [-0.763, 0.971] for C++ alternative vs C++).
659
+ 2) STDP Network: Subsequently, the Kheradpisheh et al.
660
+ network [32] is used for comparisons. This network is made of
661
+ two convolutional layers. The first layer is trained 2 times with
662
+ STDP, and the second layer is trained 20 times with STDP on
663
+ the training set. The output of the network is classified uing
664
+ the SVM classifier. The elapsed time measured consists of
665
+ the time needed to train the network on the training set and
666
+ make predictions for the testing set. The time to utilize SVM
667
+ is not taken into account because the libraries that simulate
668
+ the neural network portion are compared here. The results can
669
+ be seen in in Figure 4 and Table II.
670
+ TABLE II: Comparisons of the the runtime and accuracy
671
+ of Spyker aganist SpykeTorch on the Kheradpisheh et al.
672
+ network.
673
+ Implementation
674
+ Time
675
+ Time
676
+ (S±SD)
677
+ Accuracy
678
+ (%±SD)
679
+ Runs
680
+ SpykeTorch GPU
681
+ 47m30s
682
+ 2,850±64
683
+ 98.392±0.093
684
+ 30
685
+ Spyker GPU Single
686
+ 21m23s
687
+ 1,283±6
688
+ 98.465±0.095
689
+ 30
690
+ Spyker GPU
691
+ 05m53s
692
+ 353±9
693
+ 98.461±0.079
694
+ 30
695
+ Spyker Sparse
696
+ 08m16s
697
+ 496±1
698
+ 98.464±0.091
699
+ 30
700
+ The test results indicate that the Spyker GPU implemen-
701
+ tation is faster compared to the SpykeTorch implementation
702
+ (confidence interval [-2728, -2265]). Since the SpykeTorch
703
+ implementation processes one sample at a time, we also
704
+ implemented a single sample version on the GPU, and this
705
+ implementation runs faster compared to the SpykeTorch im-
706
+ plementation (confidence interval [-1795, -1338]). There is
707
+ also an implementation using the sparse interface of the
708
+ Spyker (that runs on CPU) that is faster than the SpykeTorch
709
+ implementation on the GPU (confidence interval [-2586, -
710
+ 2120]). These results show that the Spyker implementation is
711
+ faster while the accuracy is not significantly different ([-0.373,
712
+ 0.511] for Spyker GPU, [-0.458, 0.603] for single-sample,
713
+ and [-0.405, 0.549] for sparse implementation, all against the
714
+
715
+ 7
716
+ SpykeTorch GPU
717
+ Spyker GPU Single
718
+ Spyker GPU
719
+ Spyker CPU Sparse
720
+ 95
721
+ 96
722
+ 97
723
+ 98
724
+ 99
725
+ 100
726
+ Accuracy (%)
727
+ 98.393
728
+ 98.465
729
+ 98.462
730
+ 98.465
731
+ Accuracy results for the MNIST dataset
732
+ SpykeTorch GPU
733
+ Spyker GPU Single
734
+ Spyker GPU
735
+ Spyker CPU Sparse
736
+ 0m
737
+ 10m
738
+ 20m
739
+ 30m
740
+ 40m
741
+ 50m
742
+ Time (s)
743
+ 47m 30s
744
+ 21m 23s
745
+ 5m 53s
746
+ 8m 16s
747
+ Runtime results for the MNIST dataset
748
+ Fig. 4: Comparison plots of the runtime and accuracy of
749
+ Spyker against SpykeTorch on the Kheradpisheh et al. net-
750
+ work. The plot on the left shows shows the runtime com-
751
+ parison of Spyker and SpykeTorch implementations. The plot
752
+ on the right also compares accuracy of the two implementa-
753
+ tions. Comparisons are between GPU implementation using
754
+ SpykeTorch (SP GPU), GPU implementation using Spyker
755
+ with single-sample instead of batch processing (SP Single),
756
+ GPU implementation using Spyker (SP GPU), and Sparse CPU
757
+ implementation using Spyker (SP Sparse). The error bars are
758
+ minimum and maximum values of the samples.
759
+ SpykeTorch implementation).
760
+ B. Analyzing the Underlying Structures of the Brain
761
+ In order to demonstrate the use case and the importance
762
+ of the library in neuroscience research, a similarity analysis is
763
+ done in this section to compare the biological plausibility of an
764
+ SNN and a deep CNN model. The neural data needed for the
765
+ analysis is recorded as spiking activity and LFP signals from
766
+ Inferior Temporal (IT) cortex using a single electrode (169
767
+ sessions from two macaque monkeys, the neural data for the
768
+ monkeys are pooled together) [51]. The task implemented here
769
+ is a Rapid Serial Visual Presentation (RSVP). The intervals
770
+ are 50ms for stimulus and 450ms interstimulus. Eighy-one
771
+ greyscale images of real-world objects and Gaussian low-pass
772
+ filtered and high-pass filtered variations of some are shown
773
+ during the task (total 155 images). The categories of the stimuli
774
+ are animal face (AF), human face (HF), animal body part(AB),
775
+ human body part (HB), natual objects (N), and man-made
776
+ objects (MM).
777
+ The SNN used here is structurally similar to the one intro-
778
+ duced by Shirsavar et al. [52]. The input of the SNN is resized
779
+ to 32 and passed through 3 LoG fitlers with stds of 0.471,
780
+ 1.099, 2.042. The window sizes of the filters are 7. Then, the
781
+ output is thresholded and coded into 15 time steps. The first
782
+ convolution layer has 16 output channels with awindow size
783
+ of 5 and a padding of 2, and the second convolution layer has
784
+ 32 output channels with a window size of 3 and a padding of
785
+ 1. The pooling layers have 2 and 3 window sizes, respectively.
786
+ The layers are trained 20 times on the images, and the learning
787
+ rates are doubled after each image until they reach 0.15. Firing
788
+ times (divided by number of time steps) of the final layer is
789
+ used as the network output.
790
+ The CNN network used here is a ResNet-50 with the
791
+ classifier layer replaced. The network is not pretrained. The
792
+ input image is resized to 256 and cropped to 224. The network
793
+ is trained 15 times on the dataset with Adam optimizer and
794
+ 0.0001 learning rate. using a linear SVM classifier to classify
795
+ the 6 categories. the accuracies for the 6 classes are 51.569 ±
796
+ 2.240 (SD), 48.623 ± 2.538, and 51.247 ± 2.257 for ResNet-
797
+ 50, SNN, and an SVM classifier that is used on the average
798
+ firing rates of the neural recordings of the monkeys between
799
+ 150ms and 200m from the onset, respectively. Figure 5 Shows
800
+ the results of the analysis. The average Kendall’s Tau value
801
+ for the interval between 125ms and 175ms shown in the figure
802
+ is tested between the SNN and the ResNet. Using a Mann-
803
+ Whitney U test with the alpha value of 0.001 results in a p-
804
+ value of 2.028-07, which shows significant difference between
805
+ the two. This indicates that the SNN has a closer structure to
806
+ monkey brain.
807
+ C. Rate Coding Output
808
+ In this section, we look at the output of an SNN that uses
809
+ rate coding. The SNN network used here is the Shirsavar et
810
+ al. [52]. The number of output channels in the convolutional
811
+ layers are set to 25 and 50. The training is not changed in
812
+ that 15 time steps are used with rank order coding. However,
813
+ the inference is done with 300 time steps and rate coding.
814
+ Afterward, the spike output of 40 neurons are plotted for one
815
+ testing sample for each class shown in Figure 6. The figure also
816
+ cointains a plot of T-SNE transformed firing rates as output
817
+ fetures and the recall score for each class for the average of
818
+ 30 runs. The accuracy of the 30 runs is 95.635±0.171 on the
819
+ testing set.
820
+ IV. LIBRARY DEMONSTRATION
821
+ In this section, a sample usage of the library is illustrated.
822
+ The network used here is introduced by Shirsavar et al.
823
+ [52] to classify the MNIST dataset. The network has two
824
+ convolutional layers trained with the STDP learning rule.
825
+ The code shown in this section is only a part of the actual
826
+ implementation, with the aim of providing a simple example.
827
+ For the complete implementation, please visit the GitHub
828
+ repository of Spyker1.
829
+ 1https://github.com/ShahriarRezghi/Spyker
830
+
831
+ 8
832
+ Inferior Temporal Cortex
833
+ 450ms
834
+ Blank
835
+ 50ms
836
+ Stimulus
837
+ 450ms
838
+ Blank
839
+ Time
840
+ SNN
841
+ ResNet
842
+ +
843
+ +
844
+ Fig. 5: Similarity comparison of SNN and ResNet-50 to monkey neural data. The similarity measurement used here is the
845
+ cosine similarity. The RDM for the monkey is computed for the 50ms interval after the onset. The RDMs are adjusted with
846
+ histogram equalization. The RSA is calculated with 50ms window size and 5ms stride and 95% confidence interval. Kendall’s
847
+ Tau measurement is used for the RSA analysis. The RSA is averaged in the interval between 125ms and 175ms and compared
848
+ in the plot in the top right with 95% confidence interval.
849
+ 0
850
+ 1
851
+ 2
852
+ 3
853
+ 4
854
+ 5
855
+ 6
856
+ 7
857
+ 8
858
+ 9
859
+ Fig. 6: Raster plot of an SNN network for the MNIST test images. In this figure, 40 neurons are plotted in 300 time steps for
860
+ 10 samples of the MNIST testing set, each image belonging to one class.
861
+
862
+ 9
863
+ A. Transformation
864
+ The transformation from the input image to the network
865
+ input consists of feature enhancement and spike coding, shown
866
+ in Listing 1. Here, a module named Transform is defined that
867
+ performs the transformation when called. This module applies
868
+ 3 LoG filters with different standard deviations to the input
869
+ image with padding to keep the original width and height of
870
+ the input. The output is stored in 6 channels. Each channel
871
+ of this output is then coded into fifteen time steps using rank
872
+ order coding.
873
+ Listing 1: Implementation of the Transform module
874
+ class Transform :
875
+ def
876
+ init
877
+ ( self , device ) :
878
+ std = [0.471 , 1.099, 2.042]
879
+ self . f i l t = spyker .LoG(3 , std ,
880
+ pad=3, device=device )
881
+ def
882
+ call
883
+ ( self , data ) :
884
+ data = self . f i l t ( data )
885
+ spyker . threshold ( data , 0.01)
886
+ return spyker . code( data , 15)
887
+ B. Network
888
+ The network has two convolutional layers. Here, a module
889
+ named Network is defined (shown in Listing 2) to train the
890
+ neurons and make predictions. Here, the convolution layers
891
+ are initialized, STDP configurations are set, and the winner
892
+ selection function is wrapped with a lambda function to keep
893
+ the hyperparameters in the initialization of the function of the
894
+ network.
895
+ Listing 2: Implementation of the Network module
896
+ class Network:
897
+ def
898
+ init
899
+ ( self , device ) :
900
+ self . thresh1 , self . thresh2 = 16, 5
901
+ self .conv1 = spyker .Conv(6 , 100, 5,
902
+ pad=2, mean=.5 , std =.02, device=device )
903
+ self .conv2 = spyker .Conv(100, 200, 3,
904
+ pad=1, mean=.5 , std =.02, device=device )
905
+ config1 = spyker .STDPConfig(.0004 , −.0003)
906
+ config2 = spyker .STDPConfig(.0004 , −.0003)
907
+ self .conv1. stdpconfig = [ config1 ]
908
+ self .conv2. stdpconfig = [ config2 ]
909
+ self .wta1 = lambda x: spyker . convwta(x, 3, 5)
910
+ self .wta2 = lambda x: spyker . convwta(x, 1, 8)
911
+ C. Learning
912
+ Training each layer is done in a separate function shown in
913
+ Listing 3. The training of the layers is done in a sequantial
914
+ order (one layer after another). Training of the first layer is
915
+ done in the train layer1 function with the STDP learning rule.
916
+ Here, the output of the first convolution is computed, and
917
+ lateral inhibition is performed on it. Then, winner neurons are
918
+ selected, and STDP weight updating is performed on them.
919
+ The STDP learning rates in the first layer are multiplied by
920
+ 1.5 every 2000 samples, and the multiplying process stops
921
+ once the positive learning rate reaches 0.15. The second layer
922
+ is trained in a similar way in the train layer2 function with
923
+ the STDP learning rule.
924
+ Listing 3: The code for training of the network layers
925
+ def train layer1 ( self , data ) :
926
+ output = self .conv1( data )
927
+ spyker . threshold ( output , self . thresh1 )
928
+ spyker . inhibit ( output )
929
+ winners = self .wta1( output )
930
+ spikes = spyker . fire ( output )
931
+ self .conv1. stdp ( data , winners , spikes )
932
+ def train layer2 ( self , data ) :
933
+ data = self .conv1( data )
934
+ data = spyker . fire (data , self . thresh1 )
935
+ data = spyker . pool(data , 2)
936
+ output = self .conv2( data )
937
+ spyker . threshold ( output , self . thresh2 )
938
+ spyker . inhibit ( output )
939
+ winners = self .wta2( output )
940
+ spikes = spyker . fire ( output )
941
+ self .conv2. stdp ( data , winners , spikes )
942
+ After defining the network module, the process of training
943
+ and classification is implemented. The training process shown
944
+ in Listing 4 involves training each layer once with quantization
945
+ afterward.
946
+ Listing 4: The training process of the network
947
+ for data , target in trainset :
948
+ network . train layer1 ( transform ( data ) )
949
+ spyker . quantize (network .conv1. kernel , 0, 0.5 , 1)
950
+ for data , target in trainset :
951
+ network . train layer2 ( transform ( data ) )
952
+ spyker . quantize (network .conv2. kernel , 0, 0.5 , 1)
953
+ D. Inference
954
+ The call operator of the network shown in Listing 5 im-
955
+ plements the prediction procedure which processes the input
956
+ spikes and produces the final network output.
957
+ Listing 5: Inference function of the network
958
+ def
959
+ call
960
+ ( self , data ) :
961
+ data = self .conv1( data )
962
+ data = spyker . fire (data , self . thresh1 )
963
+ data = spyker . pool(data , 2)
964
+ data = self .conv2( data )
965
+ data = spyker . fire (data , self . thresh2 )
966
+ data = spyker . pool(data , 3)
967
+ return spyker . gather ( data ) . flatten (1)
968
+ After training, the output features for every sample in the
969
+ training set and the testing set are computed (in the gather
970
+
971
+ 10
972
+ function). Then, an SVM classifier is trained on the training
973
+ set outputs. Finally, predictions are made for the testing set
974
+ outputs (shown in Listing 6).
975
+ Listing 6: Implementation of the dimension reduction and
976
+ classification operations
977
+ xtr ,
978
+ ytr = gather ( network ,
979
+ transform ,
980
+ train )
981
+ xte , yte = gather ( network ,
982
+ transform ,
983
+ test )
984
+ svm = LinearSVC(C=2.4) . fit ( xtr ,
985
+ ytr )
986
+ pred = svm. predict ( xte )
987
+ accuracy = ( pred == testy .numpy() ) .mean()
988
+ V. DISCUSSION
989
+ Our brain has amazing capabilities. It can learn and perform
990
+ complicated tasks in a robust manner and with low power
991
+ consumption. Artificial neural networks have been created
992
+ to mimic the power of the brain processes. Deep neural
993
+ networks are ANNs that have had major success in recent
994
+ years. However, there are structural differences between these
995
+ networks and the brain, and they encounter problems when
996
+ it comes to tolerance, energy, and sample efficiency. Spiking
997
+ neural networks are the next generation of artificial neural
998
+ networks. SNNs are not a new concept. However, they have
999
+ been brought to attention recently due to their promising
1000
+ characteristics. The aim of these networks is to build a better
1001
+ model of the brain compared to DNNs.
1002
+ Several well-established simulation tools exist for DNNs.
1003
+ These tools have allowed DNNs to reach their great success
1004
+ faster and have helped them to computationally scale up. SNNs
1005
+ lack such high-performance simulation tools. There have been
1006
+ some attempts at creating such tools, but they have not been
1007
+ able to live up to expectations. In this work, we introduced
1008
+ Spyker, a high-performance library written from scratch using
1009
+ low-level tools to simulate spiking neural networks on both
1010
+ CPUs and GPUs. Despite being stand-alone, Spyker has great
1011
+ flexibility and the ability to integrate with other tools to
1012
+ create a smooth developing experience. We compared the
1013
+ performance of this library with SpykeTorch, a simulation tool
1014
+ built on the PyTorch framework. We showed that Spyker is
1015
+ multiple times faster compared to this library. Furthermore,
1016
+ to demonstrate the applicability of SNNs in describing the
1017
+ underlying neural mechanisms of the brain functions and the
1018
+ role of Spyker in this field, we compared the similarity of
1019
+ a spiking neural network implemented with this library with
1020
+ the similarity of the ResNet model to a macaque monkey
1021
+ brain. Finally, we illustrated an example implementation to
1022
+ demonstrate the easy and modern interface of the library.
1023
+ Strong SNN models can be implemented using the Spyker
1024
+ library to solve real-world machine learning problems. Fea-
1025
+ tures like fast processing and having a C++ interface alongside
1026
+ the Python interface make this library ready for both research
1027
+ and production. Generalization is an important concept in
1028
+ machine learning and having neural networks that learn and
1029
+ run fast are quite desirable. SNNs have the potential to become
1030
+ state-of-the-art models in machine learning. Other potential
1031
+ use cases of the library is to study and understand how the
1032
+ brain processes information using simulations. In other words,
1033
+ this library enables us to look at neuroscience through the eyes
1034
+ of a brain-inspired neural network.
1035
+ Although this library has been shown to be performant, there
1036
+ is room for more improvements. Spyker has a sparse interface
1037
+ that runs on the CPU. The sparse interface can be extended to
1038
+ also run on the GPU, and this can improve the performance
1039
+ even further. Furthermore, the support for a larger number
1040
+ of neural models, coding schemes, and learning rules can be
1041
+ added. This helps the library to cover a great range of SNN
1042
+ building blocks. When choosing a model to be deployed on
1043
+ embedded and neuromorphic processors, SNNs are among the
1044
+ top choices due to their energy efficiency. SNNs are often used
1045
+ in neuromorphic computing. Another direction that Spyker can
1046
+ take is in this direction. The computational efficiency of the
1047
+ sparse interface of Spyker can be further improved and made
1048
+ compatible with these types of processors.
1049
+ REFERENCES
1050
+ [1] Warren S. McCulloch and Walter Pitts. A logical calculus of the ideas
1051
+ immanent in nervous activity.
1052
+ Bulletin of Mathematical Biophysics,
1053
+ 5(4):115–133, December 1943.
1054
+ [2] F. Rosenblatt.
1055
+ The perceptron: a probabilistic model for information
1056
+ storage and organization in the brain. Psychological Review, 65(6):386–
1057
+ 408, November 1958.
1058
+ [3] J J Hopfield.
1059
+ Neural networks and physical systems with emergent
1060
+ collective computational abilities. Proceedings of the National Academy
1061
+ of Sciences of the United States of America, 79(8):2554–2558, April
1062
+ 1982.
1063
+ [4] Vinod Nair and Geoffrey E. Hinton.
1064
+ Rectified linear units improve
1065
+ restricted boltzmann machines.
1066
+ In Proceedings of the 27th Interna-
1067
+ tional Conference on International Conference on Machine Learning,
1068
+ ICML’10, pages 807–814, Madison, WI, USA, June 2010. Omnipress.
1069
+ [5] David E. Rumelhart, Geoffrey E. Hinton, and Ronald J. Williams. Learn-
1070
+ ing representations by back-propagating errors. Nature, 323(6088):533–
1071
+ 536, October 1986.
1072
+ [6] Alex Graves and J¨urgen Schmidhuber. Framewise phoneme classifica-
1073
+ tion with bidirectional LSTM and other neural network architectures.
1074
+ Neural Networks, 18(5):602–610, July 2005.
1075
+ [7] Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion
1076
+ Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. Attention
1077
+ is All you Need. In Advances in Neural Information Processing Systems,
1078
+ volume 30. Curran Associates, Inc., 2017.
1079
+ [8] Xiaohua Zhai, Alexander Kolesnikov, Neil Houlsby, and Lucas Beyer.
1080
+ Scaling Vision Transformers. pages 12104–12113, 2022.
1081
+ [9] Chen Sun, Abhinav Shrivastava, Saurabh Singh, and Abhinav Gupta.
1082
+ Revisiting Unreasonable Effectiveness of Data in Deep Learning Era.
1083
+ In 2017 IEEE International Conference on Computer Vision (ICCV),
1084
+ pages 843–852, October 2017.
1085
+ [10] D. Li, X. Chen, M. Becchi, and Z. Zong.
1086
+ Evaluating the Energy
1087
+ Efficiency of Deep Convolutional Neural Networks on CPUs and GPUs.
1088
+ In 2016 IEEE International Conferences on Big Data and Cloud
1089
+ Computing (BDCloud), Social Computing and Networking (SocialCom),
1090
+ Sustainable Computing and Communications (SustainCom) (BDCloud-
1091
+ SocialCom-SustainCom), pages 477–484, October 2016.
1092
+ [11] David B. T. McMahon and David A. Leopold.
1093
+ Stimulus timing-
1094
+ dependent plasticity in high-level vision.
1095
+ Current biology: CB,
1096
+ 22(4):332–337, February 2012.
1097
+ [12] C. Daniel Meliza and Yang Dan.
1098
+ Receptive-field modification in
1099
+ rat visual cortex induced by paired visual stimulation and single-cell
1100
+ spiking. Neuron, 49(2):183–189, January 2006.
1101
+ [13] Shiyong Huang, Carlos Rozas, Mario Trevi˜no, Jessica Contreras, Sunggu
1102
+ Yang, Lihua Song, Takashi Yoshioka, Hey-Kyoung Lee, and Alfredo
1103
+ Kirkwood. Associative Hebbian Synaptic Plasticity in Primate Visual
1104
+ Cortex. The Journal of Neuroscience, 34(22):7575–7579, May 2014.
1105
+ [14] Mohammad-Reza A. Dehaqani, Abdol-Hossein Vahabie, Roozbeh Kiani,
1106
+ Majid Nili Ahmadabadi, Babak Nadjar Araabi, and Hossein Esteky.
1107
+ Temporal dynamics of visual category representation in the macaque
1108
+ inferior temporal cortex. Journal of Neurophysiology, 116(2):587–601,
1109
+ August 2016. Publisher: American Physiological Society.
1110
+
1111
+ 11
1112
+ [15] Mohammad-Reza A Dehaqani, Abdol-Hossein Vahabie, Mohammad-
1113
+ bagher Parsa, Behrad Noudoost, and Alireza Soltani. Selective Changes
1114
+ in Noise Correlations Contribute to an Enhanced Representation of
1115
+ Saccadic Targets in Prefrontal Neuronal Ensembles. Cerebral Cortex,
1116
+ 28(8):3046–3063, August 2018.
1117
+ [16] Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Brad-
1118
+ bury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein,
1119
+ Luca Antiga, Alban Desmaison, Andreas Kopf, Edward Yang, Zachary
1120
+ DeVito, Martin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit
1121
+ Steiner, Lu Fang, Junjie Bai, and Soumith Chintala. PyTorch: An Imper-
1122
+ ative Style, High-Performance Deep Learning Library. In Advances in
1123
+ Neural Information Processing Systems, volume 32. Curran Associates,
1124
+ Inc., 2019.
1125
+ [17] TensorFlow Developers. TensorFlow, May 2022.
1126
+ [18] Tianqi Chen, Mu Li, Yutian Li, Min Lin, Naiyan Wang, Minjie Wang,
1127
+ Tianjun Xiao, Bing Xu, Chiyuan Zhang, and Zheng Zhang. MXNet:
1128
+ A Flexible and Efficient Machine Learning Library for Heterogeneous
1129
+ Distributed Systems. arXiv:1512.01274 [cs], December 2015.
1130
+ [19] Milad Mozafari, Mohammad Ganjtabesh, Abbas Nowzari-Dalini, and
1131
+ Timoth´ee Masquelier.
1132
+ SpykeTorch: Efficient Simulation of Convolu-
1133
+ tional Spiking Neural Networks With at Most One Spike per Neuron.
1134
+ Front. Neurosci., 13, 2019.
1135
+ [20] Hananel Hazan, Daniel J. Saunders, Hassaan Khan, Devdhar Patel,
1136
+ Darpan T. Sanghavi, Hava T. Siegelmann, and Robert Kozma. Bind-
1137
+ sNET: A Machine Learning-Oriented Spiking Neural Networks Library
1138
+ in Python. Front. Neuroinform., 12, 2018.
1139
+ [21] Shahriar Rezghi Shirsavar, Abdol-Hossein Vahabie, and Mohammad-
1140
+ Reza A. Dehaqani. Models Developed for Spiking Neural Networks,
1141
+ December 2022. arXiv:2212.04377 [cs, q-bio].
1142
+ [22] L. F. Abbott. Lapicque’s introduction of the integrate-and-fire model
1143
+ neuron (1907). Brain Res Bull, 50(5-6):303–304, December 1999.
1144
+ [23] Renaud Jolivet, Timothy J., and Wulfram Gerstner. The Spike Response
1145
+ Model: A Framework to Predict Neuronal Spike Trains.
1146
+ In Okyay
1147
+ Kaynak, Ethem Alpaydin, Erkki Oja, and Lei Xu, editors, Artificial
1148
+ Neural Networks and Neural Information Processing — ICANN/ICONIP
1149
+ 2003, Lecture Notes in Computer Science, pages 846–853, Berlin,
1150
+ Heidelberg, 2003. Springer.
1151
+ [24] E.M. Izhikevich. Simple model of spiking neurons. IEEE Transactions
1152
+ on Neural Networks, 14(6):1569–1572, November 2003.
1153
+ Conference
1154
+ Name: IEEE Transactions on Neural Networks.
1155
+ [25] E. D. Adrian and Y. Zotterman.
1156
+ The impulses produced by sensory
1157
+ nerve endings: Part 3. Impulses set up by Touch and Pressure. J Physiol,
1158
+ 61(4):465–483, August 1926.
1159
+ [26] Gy¨orgy Buzs´aki. Rhythms of the Brain. Oxford University Press, New
1160
+ York, 2006.
1161
+ [27] H. Markram, J. L¨ubke, M. Frotscher, and B. Sakmann.
1162
+ Regulation
1163
+ of synaptic efficacy by coincidence of postsynaptic APs and EPSPs.
1164
+ Science, 275(5297):213–215, January 1997.
1165
+ [28] W. Gerstner, R. Kempter, J. L. van Hemmen, and H. Wagner.
1166
+ A
1167
+ neuronal learning rule for sub-millisecond temporal coding.
1168
+ Nature,
1169
+ 383(6595):76–81, September 1996.
1170
+ [29] Nicolas Fr´emaux and Wulfram Gerstner.
1171
+ Neuromodulated Spike-
1172
+ Timing-Dependent Plasticity, and Theory of Three-Factor Learning
1173
+ Rules. Frontiers in Neural Circuits, 9, 2016.
1174
+ [30] Sander Bohte, Joost Kok, and Johannes Poutr´e. SpikeProp: backprop-
1175
+ agation for networks of spiking neurons. volume 48, pages 419–424,
1176
+ January 2000.
1177
+ [31] Peter U. Diehl, Daniel Neil, Jonathan Binas, Matthew Cook, Shih-Chii
1178
+ Liu, and Michael Pfeiffer. Fast-classifying, high-accuracy spiking deep
1179
+ networks through weight and threshold balancing. In 2015 International
1180
+ Joint Conference on Neural Networks (IJCNN), pages 1–8, July 2015.
1181
+ [32] Saeed Reza Kheradpisheh, Mohammad Ganjtabesh, Simon J. Thorpe,
1182
+ and Timoth´ee Masquelier.
1183
+ STDP-based spiking deep convolutional
1184
+ neural networks for object recognition.
1185
+ Neural Networks, 99:56–67,
1186
+ March 2018.
1187
+ [33] Milad Mozafari, Mohammad Ganjtabesh, Abbas Nowzari-Dalini, Si-
1188
+ mon J. Thorpe, and Timoth´ee Masquelier. Bio-inspired digit recognition
1189
+ using reward-modulated spike-timing-dependent plasticity in deep con-
1190
+ volutional networks. Pattern Recognition, 94:87–95, October 2019.
1191
+ [34] Pierre Falez, Pierre Tirilly, and Ioan Marius Bilasco. Improving STDP-
1192
+ based Visual Feature Learning with Whitening. In 2020 International
1193
+ Joint Conference on Neural Networks (IJCNN), pages 1–8, July 2020.
1194
+ ISSN: 2161-4407.
1195
+ [35] Hameda Abd El-Fattah El-Sennary, Mohamed Eid Hussien, and Abd
1196
+ El-Mgeid Amin Ali. Edge Detection of an Image Based on Extended
1197
+ Difference of Gaussian. American Journal of Computer Science and
1198
+ Technology, 2(3):35, December 2019.
1199
+ Number: 3 Publisher: Science
1200
+ Publishing Group.
1201
+ [36] R. W. Rodieck. Quantitative analysis of cat retinal ganglion cell response
1202
+ to visual stimuli. Vision Research, 5(11):583–601, December 1965.
1203
+ [37] Vincent Bonin, Valerio Mante, and Matteo Carandini.
1204
+ The suppres-
1205
+ sive field of neurons in lateral geniculate nucleus.
1206
+ The Journal of
1207
+ Neuroscience: The Official Journal of the Society for Neuroscience,
1208
+ 25(47):10844–10856, November 2005.
1209
+ [38] Tony Lindeberg.
1210
+ A computational theory of visual receptive fields.
1211
+ Biological Cybernetics, 107(6):589–635, December 2013.
1212
+ [39] B. A. Olshausen and D. J. Field. Emergence of simple-cell receptive
1213
+ field properties by learning a sparse code for natural images. Nature,
1214
+ 381(6583):607–609, June 1996.
1215
+ [40] Stefano Panzeri, Christopher D. Harvey, Eugenio Piasini, Peter E.
1216
+ Latham, and Tommaso Fellin. Cracking the Neural Code for Sensory
1217
+ Perception by Combining Statistics, Intervention, and Behavior. Neuron,
1218
+ 93(3):491–507, February 2017.
1219
+ [41] S. Thorpe, A. Delorme, and R. Van Rullen.
1220
+ Spike-based strategies
1221
+ for rapid processing.
1222
+ Neural Networks: The Official Journal of the
1223
+ International Neural Network Society, 14(6-7):715–725, 2001.
1224
+ [42] Timoth´ee Masquelier and Simon J. Thorpe.
1225
+ Unsupervised Learning
1226
+ of Visual Features through Spike Timing Dependent Plasticity. PLOS
1227
+ Computational Biology, 3(2):e31, February 2007.
1228
+ Publisher: Public
1229
+ Library of Science.
1230
+ [43] Simon Thorpe and Jacques Gautrais. Rank Order Coding. pages 113–
1231
+ 118, December 1998.
1232
+ [44] J. Gautrais and S. Thorpe. Rate coding versus temporal order coding:
1233
+ a theoretical approach. Bio Systems, 48(1-3):57–65, 1998.
1234
+ [45] Ami Citri and Robert C. Malenka. Synaptic Plasticity: Multiple Forms,
1235
+ Functions, and Mechanisms. Neuropsychopharmacology, 33(1):18–41,
1236
+ January 2008.
1237
+ [46] Olivier Bichler, Damien Querlioz, Simon J. Thorpe, Jean-Philippe Bour-
1238
+ goin, and Christian Gamrat. Extraction of temporally correlated features
1239
+ from dynamic vision sensors with spike-timing-dependent plasticity.
1240
+ Neural Networks: The Official Journal of the International Neural
1241
+ Network Society, 32:339–348, August 2012.
1242
+ [47] A. Goriely, N. Dumont, C. Dambly-Chaudi`ere, and A. Ghysen. The
1243
+ determination of sense organs in Drosophila: effect of the neuro-
1244
+ genic mutations in the embryo. Development (Cambridge, England),
1245
+ 113(4):1395–1404, December 1991.
1246
+ [48] P. Heitzler and P. Simpson. The choice of cell fate in the epidermis of
1247
+ Drosophila. Cell, 64(6):1083–1092, March 1991.
1248
+ [49] Wolfgang Maass.
1249
+ On the Computational Power of Winner-Take-All.
1250
+ Neural Computation, 12(11):2519–2535, November 2000.
1251
+ [50] Matthias Oster, Rodney Douglas, and Shih-Chii Liu. Computation with
1252
+ spikes in a winner-take-all network. Neural Computation, 21(9):2437–
1253
+ 2465, September 2009.
1254
+ [51] Esmaeil Farhang, Ramin Toosi, Behnam Karami, Roxana Koushki,
1255
+ Ehsan Rezayat, Farideh Shakerian, Jalaledin Noroozi, and Mohammad-
1256
+ Reza A. Dehaqani.
1257
+ The Effect of Spatial Frequency on the Visual
1258
+ Category Representation in the Macaque Inferior Temporal Cortex.
1259
+ bioRxiv, page 2021.12.05.470960, January 2021.
1260
+ [52] Shahriar Rezghi Shirsavar and Mohammad-Reza A. Dehaqani. A Faster
1261
+ Approach to Spiking Deep Convolutional Neural Networks, October
1262
+ 2022.
1263
+
LtFRT4oBgHgl3EQf2Di0/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
Q9E3T4oBgHgl3EQfyQvd/content/tmp_files/2301.04719v1.pdf.txt ADDED
The diff for this file is too large to render. See raw diff
 
Q9E3T4oBgHgl3EQfyQvd/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
RtAyT4oBgHgl3EQfVPc7/content/tmp_files/2301.00139v1.pdf.txt ADDED
The diff for this file is too large to render. See raw diff
 
RtAyT4oBgHgl3EQfVPc7/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
WtFPT4oBgHgl3EQfrjVl/content/tmp_files/2301.13145v1.pdf.txt ADDED
@@ -0,0 +1,820 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Accurate and efficient multiscale simulation of a
2
+ heterogeneous elastic beam via computation on
3
+ small sparse patches
4
+ A.J. Roberts∗
5
+ Thien Tran-Duc†
6
+ J.E. Bunder‡
7
+ Yannis Kevrekidis§
8
+ January 31, 2023
9
+ Abstract
10
+ Modern ‘smart’ materials have complex microscale structure, often with
11
+ unknown macroscale closure. The Equation-Free Patch Scheme empowers
12
+ us to non-intrusively, efficiently, and accurately simulate over large scales
13
+ through computations on only small well-separated patches of the microscale
14
+ system. Here the microscale system is a solid beam of random heterogeneous
15
+ elasticity. The continuing challenge is to compute the given physics on
16
+ just the microscale patches, and couple the patches across un-simulated
17
+ macroscale space, in order to establish efficiency, accuracy, consistency, and
18
+ stability on the macroscale. Dynamical systems theory supports the scheme.
19
+ This research program is to develop a systematic non-intrusive approach, both
20
+ computationally and analytically proven, to model and compute accurately
21
+ macroscale system levels of general complex physical and engineering systems.
22
+ Contents
23
+ 1
24
+ Introduction
25
+ 2
26
+ 2
27
+ Equation-free patch scheme
28
+ 4
29
+ 2.1
30
+ Scheme is non-intrusive functional ‘wrapper’ . . . . . . . . . . . . .
31
+ 4
32
+ 2.2
33
+ Scheme embeds macroscale dynamics . . . . . . . . . . . . . . . . .
34
+ 5
35
+ 3
36
+ Scheme has proven accuracy
37
+ 6
38
+ 3.1
39
+ Computation verifies exactness . . . . . . . . . . . . . . . . . . . . .
40
+ 6
41
+ 3.2
42
+ Mathematical analysis proves consistency . . . . . . . . . . . . . . .
43
+ 8
44
+ ∗School of Mathematical Sciences, University of Adelaide, South Australia.
45
+ mailto:
46
+ [email protected] https://orcid.org/0000-0001-8930-1552
47
+ †School of Mathematical Sciences, University of Adelaide, South Australia. https://orcid.
48
+ org/0000-0002-2004-5156
49
+ ‡Mathematical Sciences, University of South Australia, Australia.
50
+ https://orcid.org/
51
+ 0000-0001-5355-2288
52
+ §Departments of Chemical and Biomolecular Engineering & Applied Mathematics and
53
+ Statistics,
54
+ Johns Hopkins University,
55
+ Baltimore,
56
+ Maryland,
57
+ USA. https://orcid.org/
58
+ 0000-0003-2220-3522
59
+ 1
60
+ arXiv:2301.13145v1 [math.NA] 20 Jan 2023
61
+
62
+ 1
63
+ Introduction
64
+ 2
65
+ 4
66
+ Conclusion
67
+ 8
68
+ 1
69
+ Introduction
70
+ In structural engineering, microscale lattice materials can be light and highly stiff
71
+ with customizable macroscale mechanical properties (e.g., Somnic & Jo 2022).
72
+ The challenge we address herein is to accurately and efficiently predict macroscale
73
+ characteristics emergent from the microscale lattice. Similarly, composite materials
74
+ and structures are inherently heterogeneous and anisotropic across multiple scales.
75
+ Multiscale modelling is thus critical to the design of composite structures for
76
+ lightweight mechanical performance (e.g., Raju et al. 2021, Lucarini et al. 2021).
77
+ Such composite materials are used in electronics, space, medical, transportation,
78
+ and other industries (e.g. Matouˇs et al. 2017). Herein we establish that the Equation-
79
+ Free Patch Scheme can non-intrusively, efficiently, and accurately simulate over
80
+ macroscales through computations on only small well-separated patches of the
81
+ microscale system.
82
+ Consider an example elastic beam with heterogeneous elasticity in 2D as in Figure 1:
83
+ say 628 cm long, 20 cm wide. The beam is heterogeneous because it is constructed
84
+ from a modern material with micro-structure of size 3 cm—so that the heterogeneity
85
+ is ‘visible’ in Figure 1. With a 3 cm micro-grid, the modelling requires circa 5 000
86
+ variables. This specific scenario is easily computable, ode23 took 14 s cpu time
87
+ to simulate one period of beam bending oscillation. But if a more realistic 3 mm
88
+ micro-structure is simulated, then the computation time increases by a factor
89
+ of 1000. If 3D elasticity modelling is required for the beam, then the computation
90
+ time increases by even more orders of magnitude. The patch scheme (e.g., Samaey
91
+ et al. 2010) we develop herein potentially reduces macroscale computation time by
92
+ orders of magnitude—more reduction in higher-D space and/or smaller micro-scale.
93
+ The patch scheme achieves efficiency by only computing on small sparse patches
94
+ in space. Section 2.1 discusses how the patch scheme is non-intrusive in that it
95
+ just ‘wraps around’ a user’s microscale code—a desirable property also identified
96
+ by Biezemans et al. (2022). The patch scheme, alternatively called the gap-tooth
97
+ method, “has formal similarity with sp [superparametrization]” (Majda & Grooms
98
+ 2014, p.62) that was developed in meteorology for weather and climate predictions,
99
+ and is also akin to the so-called fe-fft and fe2 methods (Lucarini et al. 2021,
100
+ e.g.,§4.7).
101
+ Figure 1: movie of a full-domain simulation of a heterogeneous beam showing that
102
+ beam bending waves and longitudinal compression waves propagate with some
103
+ ‘average’ properties.
104
+ 0
105
+ 1
106
+ 2
107
+ 3
108
+ 4
109
+ 5
110
+ 6
111
+ space x
112
+ -0.2
113
+ 0
114
+ 0.2
115
+ y
116
+ time = 0.00, E in 0.39 3
117
+
118
+ 1
119
+ Introduction
120
+ 3
121
+ Figure 2: a small part of the
122
+ microscale grid used to code 2D
123
+ elasticity. The grid is staggered
124
+ on the microscale: ���, horizontal
125
+ displacements and velocities;
126
+ ▲, vertical displacements and
127
+ velocities; ⊚, ⊗, components of
128
+ strain and stress tensor (1).
129
+
130
+
131
+
132
+
133
+
134
+
135
+
136
+
137
+
138
+
139
+
140
+
141
+
142
+
143
+
144
+
145
+
146
+
147
+
148
+
149
+
150
+
151
+
152
+
153
+
154
+
155
+
156
+
157
+
158
+
159
+
160
+
161
+
162
+
163
+
164
+
165
+ i − 1
166
+ i
167
+ i + 1
168
+ j − 1
169
+ j
170
+ j + 1
171
+ Figure 3: example of the 2D mi-
172
+ croscale heterogeneous Young’s
173
+ modulus Eij used in computing
174
+ the elastic Lam´e parameters (3).
175
+ In this example, we choose the
176
+ heterogeneity to have microscale
177
+ period four along the beam.
178
+ <latexit sha1_base64="8j4QYKIRTLPO
179
+ mi6rtsYtpHipApY=">AkOnic7VrLbhvJFaUnrxGTMaZTY9kgXYgaSp9wMDAgMEAb
180
+ IKJsB4ZgBLMJpkS+qYr3S3LGkI/ku2yTfkR7LNLsg2H5BTV1Fi2wa9DKAaEMsdp9bVf
181
+ fce09dtThcTMq6IeSfTz760Y9/8tOfXzQ/kvfvnJrz59+utv6/lNSpejuaTefX9MK
182
+ +LSTkrXjZlMym+X1RFPh1Oiu+Gb37n7n/3tqjqcj7prlfFBfT/GpWXpajvMGl108Pz
183
+ vOvrku6+ynBTZbV5nI5g3xTgb3mfTvJnkQ9aUb346x/3j8/dqC6aWXHXOPwsnxbL6
184
+ 6IpqvnvKT394/ztKSOMrfrnw+KqnC0dfFGOmpuqWPXjxfyurFevjvuLeTlrsmnR5Nm0n
185
+ A3IGedSWMG4pVxqxU8eIPK7ATtzl7lkRBAqT7p3lD947b5Iypd1/uPnZfTLK6uZ8Ug
186
+ +XlfNYMzj1ly9vrsik+p/Lz4SQfvVmtInawrBf5qMie3T3DtXs3/yk5o27ol3CjD5j0v
187
+ p10VM3r+nSIuGTP7t3MjoRsCNRVNb+Zjb84i9OVk8nAzxNBd5kL7W8Hw3nTzKftxfv24
188
+ qS4bLz7f5XV1U5rv1G1x8uPjLT7JSgS+OMluy3FzPdBkiqn8rqf5YrCcwmqxWr6i
189
+ i+Yiq6Gz8mieTF4DkoFYydwWxKBN6UIf/GlB9AIkIRbh1DCWomBJlpR0aJYQmtqYMZ
190
+ qo0baEkpNy2MR5gSmvnZrNZSO5iVLM0mIkwTJQzu4pbSCgPDFde0hckEwxpu24wjqxze
191
+ aGKNamEqwQx3ICmN36KxlqS5dAQZwrmHacP8wDIWQSaBKPfrcEKkJ8MKynVc0K5hxO+L
192
+ Mxa8tFKbtC+auNcW7jmcwLvbmtVUpekoXftp/W2OrPfkWsMskRGXgqCsDesa64IKP9Nc
193
+ 6wgQJpx7gjLuQ4GoUqkjLoVACOJpEFyawIc1GERcigFDqDwOEwdfrWZrX2MQqFWBX2G
194
+ Epn5dY6lkERfjQA0XxDErieYtJ/iXfI2hQJaFdSUzJqwLatZ+2IQDzIUM0mP9xEg4Qm3
195
+ M3xgLqowO82nClGePu3yJuBgLKh3B0hIPYTRmOItBoI4I58ilgZHiTZpYywGAsWBSDg
196
+ c08obGBCX9hXjQJkwPqZKSBtgCIOIcWAxDpRSJpkvYS59JSi+rtIYBeKc94mrtMshB0R
197
+ umeRmDANRBs7da0IlW04GEluxDjAVy/FThcM9QO8ZQoLAaCEFhT7HGRJMLI0TcpJ
198
+ wSsmwtOaUhQF0KTnDYyiI4ByZ5u4LGrzSHAUZcSw5rbj13GiPoYyshSmFQ5DgpbGu2
199
+ 3EOpt4ioamOuzH+hxRUoUC5iYaEJHofq9wPFZNpWkiWqSNi9CYlnWZopqZJQVkamvI
200
+ dSKr1eMQkTKj0wZYO+SmpMSiYeY4DSCkVjCHE1IbRZ6aIEeCW+MQgLOglFECISJiI7A
201
+ sh27RFGvs8d3apsERkHwdL2JjlXunSRJF5qaj0ZGrXG3gQTFJKisi9oiuDyBy2y/IFD
202
+ QszRfZh5LawKu2zDvAIMRFdnXOPh8JG1gn1pKEmEi8q81D+IHBTZePXAMqrS1SL9BRg
203
+ X6hQ3VSiXamDRd5N8oVK6fDoHwKeK8VSL/FtUSqgTzoOUI0Mj/kvYwQst0Fxobfae4
204
+ zatimeMkYAKdGetcyGHEcLtk5bGQMB23DqIsF4i0MhRoKlSDjVEqxkF5GrV3PJ9fzM
205
+ RICZoM2oe6hdREXQ2HDaYWzVfpiQerKNSs6eWHafOPtANIu197GWFgcoN4LkN32LSkQ0
206
+ iYQTmivpZb4VEK81gevSpFAfMIZRLnwnlK0S2lNlSIBDQ/tANehQaBMpYpXKRDKhn7FK
207
+ tK6QBRPhKgUCGSlPyC9AwSJ92LVNnPDvOpfoNPOx+PFZN6gy65vqsuTfn2dj4tqcDnJm
208
+ 2w0r2ZFdZKNq/x24DvVk+yxE3zsB87wcdO8LETfOwEHzvBx07wsRP8/+0Es2lRX39Rz
209
+ W/rgb3oN/lwUrzCp6wuFoNRNULnt36uOjhvrsad5ej1UW2PO4f9+8O7g9G5+d9cuAe
210
+ fB4lwjCaAlaIch2vEPUw4etBxQHFXKaCPRaqG6TgPzh68B3fFIyriROKkKsR1L0IugM
211
+ kZ84BKmQxSnVB1vPhD20e040Pui1jGvPUBQRuLlLRFJoEVoaVCJtcRTrM6GJ64c4Wlzm
212
+ b2yv4ylhaCVwJqPijRQo1J3QrdUZJBInEPpKi4Lgerfl5jV6RiRVTrORL0rsWHMHiRIN
213
+ LTJKapyC6Ee6bTunQwpDYCk6Iiy827CDZfw2ILQyp1ru+x8dhmFmqAMvaAbAfo5nwx
214
+ E1EXgimDAYxRB3YDt4f/uVgt2UHoy5fXFIpNE6i23IH71ACYiSYBwg/um07F3XtO/o
215
+ lFzg0BDsNt/fhUoTjyHK9tTC7DB2ZCqWHikGlGuheIJ6mBfM+awMcT3f+l7HXNptGco
216
+ gYEOZeQe7ZgaHSwGNTN8LjkDvKMwOkt3JFP3P8WvKP2BdQZTnMpiWFraMdu0TO6lgx
217
+ IJuQa6tMLZKNxQBpCRgP9zYW26mJXci9NLTsMPHbc3cstxfQ7dM9XULsNtlrc1tsN
218
+ uh+bS7arcIbkdyP0Ut8tw0/tOwd0/AC93TdV247Dr43lLbNdYsvep65sfzl+s
219
+ n+xDBZHsrJNtbEtl7NJBtLVbETuQe+pjl+W2ix1quWn4Adq5abq3knYdnjZoavbhr
220
+ tUduMP9+9R2Q7kfirbZdjhf4fKblp+gMpumu6tsh2G25x3qOy2XafKr7snxezcfgqRh
221
+ i+1WN49efHrnK+6VbQ9oOzjqta+vXz98sn5eD6mRazZjTJ6/qVe3Z5scyrphxNCq
222
+ xg18/8tGb/Kp4haH74kh9sfSPnVfZMa6Ms8t5lblvU4SH0e9aLPNpXd9Phyd4n+bNtX
223
+ t3yBqm7nP9cPrm0lwsy9nipilmozD75c0ka+aZ+wJMNi6rYtRM7jHIR1WJDWaj67zKR0
224
+ 1RPZhpWdxMiurt9MHul4urS/dovV6dx1FdNPBkusibway4Lepm1e+DQbrJ1/bgWwahP
225
+ hN/YkdfPW+5/Lj3m95h73mP9nTvq94fel/3XvZGhz8c/vXwb4d/P/rH0b+O/n30nwD96
226
+ Elr81nvwevov/8Dw+BQBg=</latexit>
227
+ 0
228
+ 5 · 10-2 0.1
229
+ 0.15
230
+ 0.2
231
+ 0.25
232
+ -0.1
233
+ -5 · 10-2
234
+ 0
235
+ 5 · 10-2
236
+ 0.1
237
+ space x
238
+ cross-beam y
239
+ 0.5
240
+ 1
241
+ 1.5
242
+ 2
243
+ 2.5
244
+ A given microscale discretisation of heterogeneous elasticity
245
+ We adopt
246
+ a simple robust microscale approximation of 2D elasticity within the beam. On
247
+ the staggered microscale xy-grid of Figure 2 define the displacements: ▶, hori-
248
+ zontal uij(t); ▲, vertical vij(t). Microscale elasticity here first uses centred finite
249
+ differences to compute stresses, for heterogeneous Lam´e parameters λ, µ, at the
250
+ labelled microscale grid-points (Figure 2):
251
+
252
+ σxy := µij
253
+
254
+ δjuij/δyj + δivij/δxi
255
+
256
+ ;
257
+ (1a)
258
+
259
+ σxx := (λij + 2µij)δiuij/δxi + λijδjvij/δyj;
260
+ (1b)
261
+
262
+ σyy := λijδiuij/δxi + (λij + 2µij)δjvij/δyj.
263
+ (1c)
264
+ Second, centred finite differences compute the following acceleration odes
265
+
266
+ ¨uij = δiσxx/δxi + δjσxy/δyj ,
267
+ (2a)
268
+
269
+ ¨vij = δiσxy/δxi + δjσyy/δyj ,
270
+ (2b)
271
+ potentially with optional small phenomenological damping supplied by a discretisa-
272
+ tion of κ∇2 ˙uij, κ∇2 ˙vij. The patch scheme wraps around whatever microscale code
273
+ a user supplies—here it is the microscale system (1) and (2)
274
+ We nondimensionalise the system so that the density is one, and the speed of a
275
+ macroscale compression wave along the beam is about one, that is, time in these
276
+ simulations is roughly in milli-seconds.
277
+ Random periodic heterogeneity
278
+ The Lam´e parameters which appear in the
279
+ stresses (1) are
280
+ λ :=
281
+ νE
282
+ (1 + ν)(1 − 2ν),
283
+ µ :=
284
+ E
285
+ 2(1 + ν),
286
+ (3)
287
+
288
+ 2
289
+ Equation-free patch scheme
290
+ 4
291
+ in terms of Young’s modulus E and Poisson ratio ν. To have strong microscale
292
+ heterogeneity we choose these parameters randomly so that at each microscale grid-
293
+ point (iid): Eij is log-normal (here varies by factor of about ten); and νij is uniform
294
+ on [0.25, 0.35]. Figure 3 shows an example Eij. Despite such strong heterogeneity,
295
+ the movie of Figure 1 shows the macroscale dynamics appears relatively simple.
296
+ 2
297
+ Equation-free patch scheme
298
+ Instead of computing the entire beam as seen in Figure 1, the patch scheme computes
299
+ only in small sparse spatial patches such as Figure 4. In this example case, the
300
+ patch scheme reduces compute time by a factor ∝ r := (patch size)/(spacing H),
301
+ which here is just a modest factor of 1/4. But with greater scale separation and/or
302
+ in higher spatial dimensions, the scheme often reduces computational time by many
303
+ orders of magnitude.
304
+ The movie of Figure 4 shows a slow progressive wave of beam bending, together
305
+ with a not-so-slow compression wave along the beam. These macroscale predictions
306
+ are accurate (Section 3) due to the correctness of our simple coupling between
307
+ patches—even when heterogeneity is strong.
308
+ The patch scheme makes these
309
+ accurate macroscale predictions even when the macroscale closure is unknown:
310
+ the scheme does not code a closure. Further, ‘the closure’ varies depending upon
311
+ human assumptions such as choosing averaged models versus cosserat models—the
312
+ patch scheme makes no such closure assumptions. The only assumption is that the
313
+ macroscale quantities of importance vary smoothly between neighbouring patches.
314
+ 2.1
315
+ Scheme is non-intrusive functional ‘wrapper’
316
+ Consider one of the patches of the 2D beam shown in Figure 4. With the given
317
+ microscale xy-grid (Figure 2), zooming in to the microscale each patch is like that
318
+ of Figure 5. Here each patch extends across the cross-section (y-dimension) of the
319
+ beam. Open symbols in Figure 5 are ghost nodes outside the patch and implement
320
+ given stress-free top/bottom conditions on the beam. The only addition required
321
+ by the patch scheme are the edge values (‘squared’ micro-grid nodes in Figure 5)
322
+ on the left/right of each patch.
323
+ The patch scheme couples patches together by providing the patch-edge values
324
+ through interpolation across the macroscale between patches (e.g., Roberts &
325
+ Kevrekidis 2007, Roberts et al. 2014, Cao & Roberts 2016). Here we interpolate
326
+ from each of the centre patch values across the beam (i = 4 in Figure 5) of ‘nearby’
327
+ Figure 4: movie of a patch scheme simulation of a heterogeneous beam showing the
328
+ macroscale propagation across the patches of beam bending waves and longitudinal
329
+ compression waves.
330
+ 0
331
+ 1
332
+ 2
333
+ 3
334
+ 4
335
+ 5
336
+ 6
337
+ space x
338
+ -0.2
339
+ 0
340
+ 0.2
341
+ y
342
+ time = 0.00, E in 0.35 3.2
343
+
344
+ 2
345
+ Equation-free patch scheme
346
+ 5
347
+ Figure 5: one example patch
348
+ of the 2D elastic beam show-
349
+ ing the microscale staggered
350
+ grid (Figure 2). This is case
351
+ of nsubpatch = 7 micro-grid in-
352
+ tervals along the patch, and
353
+ ny = 4 intervals across the
354
+ beam.
355
+
356
+
357
+
358
+
359
+
360
+
361
+
362
+
363
+
364
+
365
+
366
+
367
+
368
+
369
+
370
+
371
+
372
+
373
+
374
+
375
+
376
+
377
+
378
+
379
+
380
+
381
+
382
+
383
+
384
+
385
+
386
+
387
+
388
+
389
+
390
+
391
+
392
+
393
+
394
+
395
+
396
+
397
+
398
+
399
+
400
+
401
+
402
+
403
+
404
+
405
+
406
+
407
+
408
+
409
+
410
+
411
+
412
+
413
+
414
+
415
+
416
+
417
+
418
+
419
+
420
+
421
+
422
+
423
+
424
+
425
+
426
+
427
+
428
+
429
+
430
+
431
+
432
+
433
+
434
+
435
+
436
+
437
+
438
+
439
+
440
+
441
+
442
+
443
+
444
+
445
+
446
+
447
+
448
+
449
+
450
+
451
+
452
+
453
+
454
+
455
+
456
+
457
+
458
+
459
+
460
+
461
+
462
+
463
+
464
+
465
+
466
+
467
+
468
+
469
+
470
+
471
+
472
+
473
+
474
+
475
+
476
+
477
+
478
+
479
+
480
+
481
+
482
+
483
+
484
+ i = 1
485
+ 2
486
+ 3
487
+ 4
488
+ 5
489
+ 6
490
+ 7
491
+ j =1
492
+ 2
493
+ 3
494
+ 4
495
+ patches, to determine the corresponding patch-edge value. Here we implement
496
+ spectral (fft) interpolation between the patches for high accuracy (Section 3).
497
+ The scheme does not presume that any average is appropriate.
498
+ This implementation shows that the patch scheme is non-intrusive (e.g., Biezemans
499
+ et al. 2022): it just ‘wraps around’ any micro-grid code a user trusts. Consequently,
500
+ we provide a toolbox (Maclean et al. 2021) for others to implement the patch
501
+ scheme around their micro-code.
502
+ 2.2
503
+ Scheme embeds macroscale dynamics
504
+ Given the patch scheme does not assume anything about what are ‘correct’
505
+ macroscale variables, a crucial question is the following: how can we be assured
506
+ that the patch scheme captures the macroscale slow dynamics?
507
+ An answer is
508
+ provided by the Whitney (1936) embedding theorem.
509
+ Roughly, the theorem is that every mD manifold is parametrisable from almost
510
+ every subspace of more than 2mD. Let’s see what this means for us. In essence,
511
+ the patch scheme provides the higher-D subspace in which the slow manifold of
512
+ the macroscale wave dynamics is embedded.
513
+ For beams in two spatial dimensions, the basic macroscale beam models have, at
514
+ each cross-section, displacement and velocity of both bending and compression.
515
+ Thus the elastic beam dynamics has a slow manifold that is m = 4D at every
516
+ cross-section.1 Alternatively, 2D cosserat beam models add a shear mode to the
517
+ macroscale model—two more variables—leading to a not-quite-so-slow manifold of
518
+ m = 6D at every cross-section. These physically based models are slow manifolds
519
+ because they focus on the relatively slow waves of solutions varying slowly in space,
520
+ and neglect all the faster high-frequency cross-waves.
521
+ In the patch scheme, Figures 1 and 4 show simulations with a cross-section of
522
+ ny = 7 micro-grid intervals, but let’s discuss the case of just ny = 4 (Figure 5).
523
+ For ny = 4, there are seven microscale nodes across each patch edge. Each node
524
+ has a displacement and velocity, and so leads to a 14D subspace for macroscale
525
+ communication between patches.
526
+ 1Such statements, invoking a manifold or subspace “at every cross-section”, are in a sense
527
+ developed by the theory of Roberts (2015). That is, in systems of large spatial extent there often
528
+ are important, spatially global, invariant manifolds of high-D that are effectively decomposable
529
+ into a union of spatially local manifolds/subspaces of relatively lower dimension—a dimension
530
+ determined by the spatial cross-section—and that are weakly coupled to neighbouring locales.
531
+
532
+ 3
533
+ Scheme has proven accuracy
534
+ 6
535
+ Figure 6: multiscale spectrum
536
+ of eigenvalues λ separates
537
+ macroscale modes on the right
538
+ from sub-patch microscale
539
+ modes on the left. The axes
540
+ are scaled nonlinearly. Here
541
+ the small viscosity is 0.001
542
+ so the microscale decays, but
543
+ the macroscale waves are long-
544
+ lasting.
545
+ -3
546
+ -1
547
+ -0.3
548
+ -0.1
549
+ -0.03
550
+ -0.01
551
+ 0
552
+ -100
553
+ -30
554
+ -10
555
+ -3
556
+ -1
557
+ -0.3
558
+ -0.10
559
+ 0.1
560
+ 0.3
561
+ 1
562
+ 3
563
+ 10
564
+ 30
565
+ 100
566
+ ℜλ
567
+ ℑλ
568
+ Because 14 > 2 · 6 > 2 · 4 , the Whitney embedding theorem asserts that the
569
+ patch scheme exchanges enough information to almost surely parametrise both
570
+ such slow manifolds of the macroscale dynamics. The patch scheme does not need
571
+ to explicitly compute and exchange specific assumed macroscale average quantities.
572
+ 3
573
+ Scheme has proven accuracy
574
+ Section 3.2 discusses established theory which generally proves that the patch
575
+ scheme makes accurate macroscale predictions. Such proofs are in stark contrast
576
+ to the vast machine learning/artificial intelligence developments which prove
577
+ very few general results: for example, Brenner & Koumoutsakos (2021) comment
578
+ “. . . ml studies, as the lack of rigorous theory does not offer (yet!) guarantees
579
+ of convergence”. Before discussing theory, we first report some computational
580
+ verification of high accuracy.
581
+ 3.1
582
+ Computation verifies exactness
583
+ Here we restricted attention to linear elasticity so we know that the wrapped
584
+ patch system is fully characterised by the resultant Jacobian matrix. We numeri-
585
+ cally compute the Jacobian matrix of the patch scheme by elementary numerical
586
+ differentiation.
587
+ Because of the macroscale translational invariance of the patch scheme, the
588
+ macroscale eigenvectors are correctly sinusoidal. Hence the only macroscale er-
589
+ rors occur in the eigenvalues of the Jacobian. Figure 6 plots the spectrum of all
590
+ eigenvalues for one example of random heterogeneity, in the case of five patches for
591
+ simplicity. Observe there are:
592
+ • (on the right) four λ = 0 of rigid beam motion;
593
+ • four −0.001 ± i 1.057 and four −0.003 ± i 2.111 of compressions waves;
594
+ • four −0.001 ± i 0.061 and four −0.004 ± i 0.237 of beam bending waves;
595
+ • with the above macroscale eigenvalues separated by a spectral gap from the
596
+ following sub-patch microscale eigenvalues;
597
+
598
+ 3
599
+ Scheme has proven accuracy
600
+ 7
601
+ Table 1: error in patch scheme’s
602
+ macroscale eigenvalues λ for
603
+ various patch size ratios r: the
604
+ macroscale λs are exact to round-
605
+ off error—due to patch coupling by
606
+ spectral interpolation.
607
+ macro-eigenvalue
608
+ r = 1
609
+ 2
610
+ r = 1
611
+ 4
612
+ r = 1
613
+ 8
614
+ −0.001 ± i 0.061
615
+ 2e-12
616
+ 1e-12
617
+ 2e-13
618
+ −0.001 ± i 0.061
619
+ 2e-12
620
+ 4e-12
621
+ 2e-12
622
+ −0.004 ± i 0.237
623
+ 1e-12
624
+ 8e-13
625
+ 3e-12
626
+ −0.004 ± i 0.237
627
+ 1e-12
628
+ 2e-12
629
+ 3e-12
630
+ −0.001 ± i 1.057
631
+ 7e-13
632
+ 4e-13
633
+ 6e-13
634
+ −0.001 ± i 1.057
635
+ 6e-13
636
+ 5e-13
637
+ 6e-13
638
+ −0.003 ± i 2.111
639
+ 1e-13
640
+ 2e-13
641
+ 2e-13
642
+ −0.003 ± i 2.111
643
+ 4e-13
644
+ 5e-13
645
+ 2e-13
646
+ Figure 7: multiscale spectrum
647
+ of eigenvalues λ for the patch
648
+ scheme in the case of zero viscos-
649
+ ity. The horizontal axis shows
650
+ that all modes have zero real-
651
+ part to numerical round-off er-
652
+ ror. That is, in the case of zero
653
+ viscosity, this patch scheme pre-
654
+ serves the wave nature of the
655
+ underlying physics.
656
+ -5e-13
657
+ -2e-13
658
+ -1e-130
659
+ 1e-13
660
+ 2e-13
661
+ 5e-13
662
+ -100
663
+ -30
664
+ -10
665
+ -3
666
+ -1
667
+ -0.3
668
+ -0.10
669
+ 0.1
670
+ 0.3
671
+ 1
672
+ 3
673
+ 10
674
+ 30
675
+ 100
676
+ ℜλ
677
+ ℑλ
678
+ • (on the left) many ℜλ < −0.1 of uninteresting sub-patch micro-scale fast-
679
+ waves (headed by ten eigenvalues around −0.14 ± i 9.29).
680
+ To quantify the accuracy, Table 1 compares eigenvalues obtained from full-domain
681
+ code, with the above macroscale eigenvalues obtained by the wrapped patch scheme.
682
+ For all patch size ratios and heterogeneities tested, the patch scheme’s macroscale
683
+ eigenvalues are exact to numerical round-off error.
684
+ Such exactness is due to the spectral interpolation used here. If, instead of spectral,
685
+ local polynomial interpolation of degree p is used to couple the patches, then
686
+ generally the patch scheme has macroscale errors ∝ Hp where H = inter-patch
687
+ spacing (e.g., Roberts & Kevrekidis 2007, Roberts et al. 2014).
688
+ Undamped waves?
689
+ With zero viscosity, there are only oscillations in the under-
690
+ lying physics. In such a scenario computational methods are very delicate. Here,
691
+ Figure 7 illustrates that all eigenvalues of the patch scheme have |ℜλ| < 10−12.2
692
+ Hence, even with no viscosity, the patch scheme preserves the oscillatory wave
693
+ nature of the heterogeneous physics.
694
+ There is a perception that the patch scheme “only works well on problems with
695
+ an inertial manifold and for systems in which most modes are strongly decaying”
696
+ (Majda & Grooms 2014, p.62). This verification of accuracy for purely elastic
697
+ 2In some realisations of the heterogeneity, the sensitive multiplicity four eigenvalue λ = 0
698
+ numerically splits into four showing |ℜλ| up to 10−6 due to round-off errors.
699
+
700
+ 4
701
+ Conclusion
702
+ 8
703
+ beams shows that this perception is false. Applications and theory for other wave
704
+ systems also refute this perception (e.g., Cao & Roberts 2016, Bunder et al. 2021,
705
+ Divahar et al. 2022).
706
+ 3.2
707
+ Mathematical analysis proves consistency
708
+ Mathematical analysis has proven properties of the patch scheme in general. Mostly,
709
+ the published proofs explicitly address dissipative (nonlinear) systems. However,
710
+ as discussed by Bunder et al. (2021), the patch scheme in space only recasts spatial
711
+ interactions, so whether the time derivative is ∂/∂t of dissipation or ∂2/∂t2 of
712
+ waves makes little difference.
713
+ Two complementary types of results have been proven. They involve the spacing
714
+ between patch centres H.
715
+ First, Centre Manifold Theory may be applied at
716
+ finite spacing H by introducing a ‘bookkeeping’ parameter γ to label inter-patch
717
+ communication (e.g., Roberts et al. 2014, §2) to prove the existence of a slow
718
+ manifold in the patch scheme (including when it is applied to nonlinear systems).
719
+ Then the parameter γ structures inter-patch interactions, and their algebraic
720
+ expression, to empower theory based at γ = 0, via regular perturbation, to address
721
+ finite γ such as the case of full coupling γ = 1 (e.g., Roberts et al. 2014, Cor. 2).
722
+ Second, the patch scheme is consistent with the underlying micro-code as the
723
+ patch spacing H → 0 (e.g., Roberts et al. 2014, Thm. 7). The consistency is
724
+ that the macroscale of the patch scheme is the same as the macroscale of the
725
+ given micro-coded system, to errors O
726
+
727
+ Hp�
728
+ when using polynomial interpolation
729
+ of degree p. For example, spectral interpolation corresponds to ‘p = ∞’ so then
730
+ errors vanish to all orders as in Table 1.
731
+ These results and general proofs were first done for homogeneous systems (e.g.,
732
+ Roberts & Kevrekidis 2007, Roberts et al. 2014). They were subsequently ex-
733
+ tended to heterogeneous microscales (Bunder et al. 2017), and recently extended
734
+ to alternative inter-patch coupling that preserves self-adjointness (Bunder et al.
735
+ 2021). Interestingly, the extension of the theoretical support to heterogeneous
736
+ cases invokes the ensemble of all phase-shifts of the heterogeneity. The ensemble is
737
+ spatially homogeneous, so the homogeneous proofs and results apply to establish
738
+ the heterogeneous results.
739
+ 4
740
+ Conclusion
741
+ As an initial exploration of the patch scheme for homogenisation of heterogeneous
742
+ elasticity, we considered the prototypical case of a 2D elastic beam. The scheme
743
+ gives a non-intrusive and efficient computational homogenisation of given microscale
744
+ system via spatially sparse small patches. The patch coupling has proven accuracy,
745
+ controllable error, at finite scale separation.
746
+ The patch scheme makes only one assumption: in the scenarios of interest to a
747
+ user, there is no significant spatial structures in the mesoscale between the patch
748
+ spacing H and the microscale resolved in the patches. In contrast to most other
749
+ multiscale methods, there is: no assumed boundary conditions on Representative
750
+ Volume Elements (variously periodic, Dirichlet, Neumann); no explicitly assuming
751
+
752
+ References
753
+ 9
754
+ slow variables; and no presumed necessary variational principle. The scheme is
755
+ entirely physically interpretable: there is no hidden mystic machinations of neural
756
+ networks (e.g., Brenner & Koumoutsakos 2021)
757
+ The patch scheme is simple to apply. In contrast to other multiscale methods
758
+ there is: no arbitrary averaging; no oversampling regions; no buffer regions; no
759
+ action regions; no guessed fast/slow variables; no epsilons; and no limits. As a
760
+ non-intrusive ‘wrapper’, anyone can start using the patch scheme via a Matlab/
761
+ Octave Toolbox (Maclean et al. 2021, Roberts et al. 2019���2023)
762
+ Acknowledgements
763
+ This research was supported by Australian Research Coun-
764
+ cil grants DP220103156 and DP200103097.
765
+ References
766
+ Biezemans, R. A., Le Bris, C., Legoll, F. & Lozinski, A. (2022), Non-intrusive
767
+ implementation of a wide variety of Multiscale Finite Element Methods, Technical
768
+ report, http://arxiv.org/abs/2211.17024.
769
+ Brenner, M. P. & Koumoutsakos, P. (2021), ‘Editorial: Machine Learning and
770
+ Physical Review Fluids: An Editorial Perspective’, Physical Review Fluids
771
+ 6(7), 070001.
772
+ Bunder, J. E., Kevrekidis, I. G. & Roberts, A. J. (2021), ‘Equation-free patch
773
+ scheme for efficient computational homogenisation via self-adjoint coupling’,
774
+ Numerische Mathematik 149(2), 229–272.
775
+ Bunder, J. E., Roberts, A. J. & Kevrekidis, I. G. (2017), ‘Good coupling for the mul-
776
+ tiscale patch scheme on systems with microscale heterogeneity’, J. Computational
777
+ Physics 337, 154–174.
778
+ Cao, M. & Roberts, A. J. (2016), ‘Multiscale modelling couples patches of nonlinear
779
+ wave-like simulations’, IMA J. Applied Maths. 81(2), 228–254.
780
+ Divahar, J., Roberts, A. J., Mattner, T. W., Bunder, J. E. & Kevrekidis, I. G.
781
+ (2022), Two novel families of multiscale staggered patch schemes efficiently
782
+ simulate large-scale, weakly damped, linear waves, Technical report, https:
783
+ //arxiv.org/abs/2210.15823.
784
+ Lucarini, S., Upadhyay, M. V. & Segurado, J. (2021), ‘FFT based approaches
785
+ in micromechanics: fundamentals, methods and applications’, Modelling and
786
+ Simulation in Materials Science and Engineering 30(2), 023002.
787
+ Maclean, J., Bunder, J. E. & Roberts, A. J. (2021), ‘A toolbox of equation-free
788
+ functions in matlab/octave for efficient system level simulation’, Numerical
789
+ Algorithms 87, 1729–1748.
790
+ Majda, A. J. & Grooms, I. (2014), ‘New perspectives on superparameterization for
791
+ geophysical turbulence’, Journal of Computational Physics 271, 60–77.
792
+ Matouˇs, K., Geers, M. G. D., Kouznetsova, V. G. & Gillman, A. (2017), ‘A review of
793
+ predictive nonlinear theories for multiscale modeling of heterogeneous materials’,
794
+ Journal of Computational Physics 330, 192–220.
795
+
796
+ References
797
+ 10
798
+ Raju, K., Tay, T.-E. & Tan, V. B. C. (2021), ‘A review of the FE2 method for
799
+ composites’, Multiscale and Multidisciplinary Modeling, Experiments and Design
800
+ 4, 1–24.
801
+ Roberts, A. J. (2015), ‘Macroscale, slowly varying, models emerge from the mi-
802
+ croscale dynamics in long thin domains’, IMA Journal of Applied Mathematics
803
+ 80(5), 1492–1518.
804
+ Roberts, A. J. & Kevrekidis, I. G. (2007), ‘General tooth boundary conditions for
805
+ equation free modelling’, SIAM J. Scientific Computing 29(4), 1495–1510.
806
+ Roberts, A. J., MacKenzie, T. & Bunder, J. (2014), ‘A dynamical systems approach
807
+ to simulating macroscale spatial dynamics in multiple dimensions’, J. Engineering
808
+ Mathematics 86(1), 175–207.
809
+ Roberts, A. J., Maclean, J. & Bunder, J. E. (2019–2023), Equation-free function tool-
810
+ box for matlab/octave, Technical report, [https://github.com/uoa1184615/
811
+ EquationFreeGit].
812
+ Samaey, G., Roberts, A. J. & Kevrekidis, I. G. (2010), Equation-free computation:
813
+ an overview of patch dynamics, in J. Fish, ed., ‘Multiscale methods: bridging the
814
+ scales in science and engineering’, Oxford University Press, chapter 8, pp. 216–
815
+ 246.
816
+ Somnic, J. & Jo, B. W. (2022), ‘Status and challenges in homogenization methods
817
+ for lattice materials’, Materials 15(2), 605.
818
+ Whitney, H. (1936), ‘Differentiable manifolds’, Annals of Mathematics 37(3), 645–
819
+ 680.
820
+
WtFPT4oBgHgl3EQfrjVl/content/tmp_files/load_file.txt ADDED
@@ -0,0 +1,516 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ filepath=/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf,len=515
2
+ page_content='Accurate and efficient multiscale simulation of a heterogeneous elastic beam via computation on small sparse patches A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
3
+ page_content='J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
4
+ page_content=' Roberts∗ Thien Tran-Duc† J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
5
+ page_content='E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
6
+ page_content=' Bunder‡ Yannis Kevrekidis§ January 31, 2023 Abstract Modern ‘smart’ materials have complex microscale structure, often with unknown macroscale closure.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
7
+ page_content=' The Equation-Free Patch Scheme empowers us to non-intrusively, efficiently, and accurately simulate over large scales through computations on only small well-separated patches of the microscale system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
8
+ page_content=' Here the microscale system is a solid beam of random heterogeneous elasticity.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
9
+ page_content=' The continuing challenge is to compute the given physics on just the microscale patches, and couple the patches across un-simulated macroscale space, in order to establish efficiency, accuracy, consistency, and stability on the macroscale.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
10
+ page_content=' Dynamical systems theory supports the scheme.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
11
+ page_content=' This research program is to develop a systematic non-intrusive approach, both computationally and analytically proven, to model and compute accurately macroscale system levels of general complex physical and engineering systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
12
+ page_content=' Contents 1 Introduction 2 2 Equation-free patch scheme 4 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
13
+ page_content='1 Scheme is non-intrusive functional ‘wrapper’ .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
14
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
15
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
16
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
17
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
18
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
19
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
20
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
21
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
22
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
23
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
24
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
25
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
26
+ page_content=' 4 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
27
+ page_content='2 Scheme embeds macroscale dynamics .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
28
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
29
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
30
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
31
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
32
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
33
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
34
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
35
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
36
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
37
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
38
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
39
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
40
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
41
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
42
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
43
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
44
+ page_content=' 5 3 Scheme has proven accuracy 6 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
45
+ page_content='1 Computation verifies exactness .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
46
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
47
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
48
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
49
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
50
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
51
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
52
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
53
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
54
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
55
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
56
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
57
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
58
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
59
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
60
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
61
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
62
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
63
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
64
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
65
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
66
+ page_content=' 6 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
67
+ page_content='2 Mathematical analysis proves consistency .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
68
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
69
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
70
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
71
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
72
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
73
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
74
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
75
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
76
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
77
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
78
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
79
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
80
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
81
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
82
+ page_content=' 8 ∗School of Mathematical Sciences, University of Adelaide, South Australia.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
83
+ page_content=' mailto: ProfAJRoberts@protonmail.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
84
+ page_content='com https://orcid.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
85
+ page_content='org/0000-0001-8930-1552 †School of Mathematical Sciences, University of Adelaide, South Australia.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
86
+ page_content=' https://orcid.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
87
+ page_content=' org/0000-0002-2004-5156 ‡Mathematical Sciences, University of South Australia, Australia.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
88
+ page_content=' https://orcid.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
89
+ page_content='org/ 0000-0001-5355-2288 §Departments of Chemical and Biomolecular Engineering & Applied Mathematics and Statistics, Johns Hopkins University, Baltimore, Maryland, USA.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
90
+ page_content=' https://orcid.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
91
+ page_content='org/ 0000-0003-2220-3522 1 arXiv:2301.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
92
+ page_content='13145v1 [math.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
93
+ page_content='NA] 20 Jan 2023 1 Introduction 2 4 Conclusion 8 1 Introduction In structural engineering, microscale lattice materials can be light and highly stiff with customizable macroscale mechanical properties (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
94
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
95
+ page_content=', Somnic & Jo 2022).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
96
+ page_content=' The challenge we address herein is to accurately and efficiently predict macroscale characteristics emergent from the microscale lattice.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
97
+ page_content=' Similarly, composite materials and structures are inherently heterogeneous and anisotropic across multiple scales.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
98
+ page_content=' Multiscale modelling is thus critical to the design of composite structures for lightweight mechanical performance (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
99
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
100
+ page_content=', Raju et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
101
+ page_content=' 2021, Lucarini et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
102
+ page_content=' 2021).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
103
+ page_content=' Such composite materials are used in electronics, space, medical, transportation, and other industries (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
104
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
105
+ page_content=' Matouˇs et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
106
+ page_content=' 2017).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
107
+ page_content=' Herein we establish that the Equation- Free Patch Scheme can non-intrusively, efficiently, and accurately simulate over macroscales through computations on only small well-separated patches of the microscale system.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
108
+ page_content=' Consider an example elastic beam with heterogeneous elasticity in 2D as in Figure 1: say 628 cm long, 20 cm wide.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
109
+ page_content=' The beam is heterogeneous because it is constructed from a modern material with micro-structure of size 3 cm—so that the heterogeneity is ‘visible’ in Figure 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
110
+ page_content=' With a 3 cm micro-grid, the modelling requires circa 5 000 variables.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
111
+ page_content=' This specific scenario is easily computable, ode23 took 14 s cpu time to simulate one period of beam bending oscillation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
112
+ page_content=' But if a more realistic 3 mm micro-structure is simulated, then the computation time increases by a factor of 1000.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
113
+ page_content=' If 3D elasticity modelling is required for the beam, then the computation time increases by even more orders of magnitude.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
114
+ page_content=' The patch scheme (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
115
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
116
+ page_content=', Samaey et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
117
+ page_content=' 2010) we develop herein potentially reduces macroscale computation time by orders of magnitude—more reduction in higher-D space and/or smaller micro-scale.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
118
+ page_content=' The patch scheme achieves efficiency by only computing on small sparse patches in space.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
119
+ page_content=' Section 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
120
+ page_content='1 discusses how the patch scheme is non-intrusive in that it just ‘wraps around’ a user’s microscale code—a desirable property also identified by Biezemans et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
121
+ page_content=' (2022).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
122
+ page_content=' The patch scheme, alternatively called the gap-tooth method, “has formal similarity with sp [superparametrization]” (Majda & Grooms 2014, p.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
123
+ page_content='62) that was developed in meteorology for weather and climate predictions, and is also akin to the so-called fe-fft and fe2 methods (Lucarini et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
124
+ page_content=' 2021, e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
125
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
126
+ page_content=',§4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
127
+ page_content='7).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
128
+ page_content=' Figure 1: movie of a full-domain simulation of a heterogeneous beam showing that beam bending waves and longitudinal compression waves propagate with some ‘average’ properties.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
129
+ page_content=' 0 1 2 3 4 5 6 space x 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
130
+ page_content='2 0 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
131
+ page_content='2 y time = 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
132
+ page_content='00, E in 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
133
+ page_content='39 3 1 Introduction 3 Figure 2: a small part of the microscale grid used to code 2D elasticity.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
134
+ page_content=' The grid is staggered on the microscale: ▶, horizontal displacements and velocities;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
135
+ page_content=' ▲, vertical displacements and velocities;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
136
+ page_content=' ⊚, ⊗, components of strain and stress tensor (1).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
137
+ page_content=' ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ i − 1 i i + 1 j − 1 j j + 1 Figure 3: example of the 2D mi- croscale heterogeneous Young’s modulus Eij used in computing the elastic Lam´e parameters (3).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
138
+ page_content=' In this example, we choose the heterogeneity to have microscale period four along the beam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
139
+ page_content=' ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
140
+ page_content='<latexit sha1_base64="8j4QYKIRTLPO ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
141
+ page_content='mi6rtsYtpHipApY=">AkOnic7VrLbhvJFaUnrxGTMaZTY9kgXYgaSp9wMDAgMEAb ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
142
+ page_content='IKJsB4ZgBLMJpkS+qYr3S3LGkI/ku2yTfkR7LNLsg2H5BTV1Fi2wa9DKAaEMsdp9bVf ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
143
+ page_content='fce09dtThcTMq6IeSfTz760Y9/8tOfXzQ/kvfvnJrz59+utv6/lNSpejuaTefX9MK ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
144
+ page_content='+LSTkrXjZlMym+X1RFPh1Oiu+Gb37n7n/3tqjqcj7prlfFBfT/GpWXpajvMGl108Pz ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
145
+ page_content='vOvrku6+ynBTZbV5nI5g3xTgb3mfTvJnkQ9aUb346x/3j8/dqC6aWXHXOPwsnxbL6 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
146
+ page_content='6IpqvnvKT394/ztKSOMrfrnw+KqnC0dfFGOmpuqWPXjxfyurFevjvuLeTlrsmnR5Nm0n ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
147
+ page_content='A3IGedSWMG4pVxqxU8eIPK7ATtzl7lkRBAqT7p3lD947b5Iypd1/uPnZfTLK6uZ8Ug ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
148
+ page_content='+XlfNYMzj1ly9vrsik+p/Lz4SQfvVmtInawrBf5qMie3T3DtXs3/yk5o27ol3CjD5j0v ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
149
+ page_content='p10VM3r+nSIuGTP7t3MjoRsCNRVNb+Zjb84i9OVk8nAzxNBd5kL7W8Hw3nTzKftxfv24 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
150
+ page_content='qS4bLz7f5XV1U5rv1G1x8uPjLT7JSgS+OMluy3FzPdBkiqn8rqf5YrCcwmqxWr6i ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
151
+ page_content='i+Yiq6Gz8mieTF4DkoFYydwWxKBN6UIf/GlB9AIkIRbh1DCWomBJlpR0aJYQmtqYMZ ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
152
+ page_content='qo0baEkpNy2MR5gSmvnZrNZSO5iVLM0mIkwTJQzu4pbSCgPDFde0hckEwxpu24wjqxze ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
153
+ page_content='aGKNamEqwQx3ICmN36KxlqS5dAQZwrmHacP8wDIWQSaBKPfrcEKkJ8MKynVc0K5hxO+L ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
154
+ page_content='Mxa8tFKbtC+auNcW7jmcwLvbmtVUpekoXftp/W2OrPfkWsMskRGXgqCsDesa64IKP9Nc ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
155
+ page_content='6wgQJpx7gjLuQ4GoUqkjLoVACOJpEFyawIc1GERcigFDqDwOEwdfrWZrX2MQqFWBX2G ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
156
+ page_content='Epn5dY6lkERfjQA0XxDErieYtJ/iXfI2hQJaFdSUzJqwLatZ+2IQDzIUM0mP9xEg4Qm3 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
157
+ page_content='M3xgLqowO82nClGePu3yJuBgLKh3B0hIPYTRmOItBoI4I58ilgZHiTZpYywGAsWBSDg ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
158
+ page_content='c08obGBCX9hXjQJkwPqZKSBtgCIOIcWAxDpRSJpkvYS59JSi+rtIYBeKc94mrtMshB0R ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
159
+ page_content='umeRmDANRBs7da0IlW04GEluxDjAVy/FThcM9QO8ZQoLAaCEFhT7HGRJMLI0TcpJ ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
160
+ page_content='wSsmwtOaUhQF0KTnDYyiI4ByZ5u4LGrzSHAUZcSw5rbj13GiPoYyshSmFQ5DgpbGu2 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
161
+ page_content='3EOpt4ioamOuzH+hxRUoUC5iYaEJHofq9wPFZNpWkiWqSNi9CYlnWZopqZJQVkamvI ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
162
+ page_content='dSKr1eMQkTKj0wZYO+SmpMSiYeY4DSCkVjCHE1IbRZ6aIEeCW+MQgLOglFECISJiI7A ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
163
+ page_content='sh27RFGvs8d3apsERkHwdL2JjlXunSRJF5qaj0ZGrXG3gQTFJKisi9oiuDyBy2y/IFD ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
164
+ page_content='QszRfZh5LawKu2zDvAIMRFdnXOPh8JG1gn1pKEmEi8q81D+IHBTZePXAMqrS1SL9BRg ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
165
+ page_content='X6hQ3VSiXamDRd5N8oVK6fDoHwKeK8VSL/FtUSqgTzoOUI0Mj/kvYwQst0Fxobfae4 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
166
+ page_content='zatimeMkYAKdGetcyGHEcLtk5bGQMB23DqIsF4i0MhRoKlSDjVEqxkF5GrV3PJ9fzM ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
167
+ page_content='RICZoM2oe6hdREXQ2HDaYWzVfpiQerKNSs6eWHafOPtANIu197GWFgcoN4LkN32LSkQ0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
168
+ page_content='iYQTmivpZb4VEK81gevSpFAfMIZRLnwnlK0S2lNlSIBDQ/tANehQaBMpYpXKRDKhn7FK ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
169
+ page_content='tK6QBRPhKgUCGSlPyC9AwSJ92LVNnPDvOpfoNPOx+PFZN6gy65vqsuTfn2dj4tqcDnJm ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
170
+ page_content='2w0r2ZFdZKNq/x24DvVk+yxE3zsB87wcdO8LETfOwEHzvBx07wsRP8/+0Es2lRX39Rz ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
171
+ page_content='W/rgb3oN/lwUrzCp6wuFoNRNULnt36uOjhvrsad5ej1UW2PO4f9+8O7g9G5+d9cuAe ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
172
+ page_content='fB4lwjCaAlaIch2vEPUw4etBxQHFXKaCPRaqG6TgPzh68B3fFIyriROKkKsR1L0IugM ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
173
+ page_content='kZ84BKmQxSnVB1vPhD20e040Pui1jGvPUBQRuLlLRFJoEVoaVCJtcRTrM6GJ64c4Wlzm ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
174
+ page_content='b2yv4ylhaCVwJqPijRQo1J3QrdUZJBInEPpKi4Lgerfl5jV6RiRVTrORL0rsWHMHiRIN ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
175
+ page_content='LTJKapyC6Ee6bTunQwpDYCk6Iiy827CDZfw2ILQyp1ru+x8dhmFmqAMvaAbAfo5nwx ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
176
+ page_content='E1EXgimDAYxRB3YDt4f/uVgt2UHoy5fXFIpNE6i23IH71ACYiSYBwg/um07F3XtO/o ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
177
+ page_content='lFzg0BDsNt/fhUoTjyHK9tTC7DB2ZCqWHikGlGuheIJ6mBfM+awMcT3f+l7HXNptGco ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
178
+ page_content='gYEOZeQe7ZgaHSwGNTN8LjkDvKMwOkt3JFP3P8WvKP2BdQZTnMpiWFraMdu0TO6lgx ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
179
+ page_content='IJuQa6tMLZKNxQBpCRgP9zYW26mJXci9NLTsMPHbc3cstxfQ7dM9XULsNtlrc1tsN ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
180
+ page_content='uh+bS7arcIbkdyP0Ut8tw0/tOwd0/AC93TdV247Dr43lLbNdYsvep65sfzl+s ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
181
+ page_content='n+xDBZHsrJNtbEtl7NJBtLVbETuQe+pjl+W2ix1quWn4Adq5abq3knYdnjZoavbhr ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
182
+ page_content='tUduMP9+9R2Q7kfirbZdjhf4fKblp+gMpumu6tsh2G25x3qOy2XafKr7snxezcfgqRh ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
183
+ page_content='i+1WN49efHrnK+6VbQ9oOzjqta+vXz98sn5eD6mRazZjTJ6/qVe3Z5scyrphxNCq ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
184
+ page_content='xg18/8tGb/Kp4haH74kh9sfSPnVfZMa6Ms8t5lblvU4SH0e9aLPNpXd9Phyd4n+bNtX ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
185
+ page_content='t3yBqm7nP9cPrm0lwsy9nipilmozD75c0ka+aZ+wJMNi6rYtRM7jHIR1WJDWaj67zKR0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
186
+ page_content='1RPZhpWdxMiurt9MHul4urS/dovV6dx1FdNPBkusibway4Lepm1e+DQbrJ1/bgWwahP ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
187
+ page_content='hN/YkdfPW+5/Lj3m95h73mP9nTvq94fel/3XvZGhz8c/vXwb4d/P/rH0b+O/n30nwD96 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
188
+ page_content='Elr81nvwevov/8Dw+BQBg=</latexit> ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
189
+ page_content='0 ' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
190
+ page_content='5 · 10-2 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
191
+ page_content='1 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
192
+ page_content='15 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
193
+ page_content='2 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
194
+ page_content='25 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
195
+ page_content='1 5 · 10-2 0 5 · 10-2 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
196
+ page_content='1 space x cross-beam y 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
197
+ page_content='5 1 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
198
+ page_content='5 2 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
199
+ page_content='5 A given microscale discretisation of heterogeneous elasticity We adopt a simple robust microscale approximation of 2D elasticity within the beam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
200
+ page_content=' On the staggered microscale xy-grid of Figure 2 define the displacements: ▶, hori- zontal uij(t);' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
201
+ page_content=' ▲, vertical vij(t).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
202
+ page_content=' Microscale elasticity here first uses centred finite differences to compute stresses, for heterogeneous Lam´e parameters λ, µ, at the labelled microscale grid-points (Figure 2): ⊗ σxy := µij � δjuij/δyj + δivij/δxi � ;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
203
+ page_content=' (1a) ⊚ σxx := (λij + 2µij)δiuij/δxi + λijδjvij/δyj;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
204
+ page_content=' (1b) ⊚ σyy := λijδiuij/δxi + (λij + 2µij)δjvij/δyj.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
205
+ page_content=' (1c) Second, centred finite differences compute the following acceleration odes ▶ ¨uij = δiσxx/δxi + δjσxy/δyj , (2a) ▲ ¨vij = δiσxy/δxi + δjσyy/δyj , (2b) potentially with optional small phenomenological damping supplied by a discretisa- tion of κ∇2 ˙uij, κ∇2 ˙vij.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
206
+ page_content=' The patch scheme wraps around whatever microscale code a user supplies—here it is the microscale system (1) and (2) We nondimensionalise the system so that the density is one, and the speed of a macroscale compression wave along the beam is about one, that is, time in these simulations is roughly in milli-seconds.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
207
+ page_content=' Random periodic heterogeneity The Lam´e parameters which appear in the stresses (1) are λ := νE (1 + ν)(1 − 2ν), µ := E 2(1 + ν), (3) 2 Equation-free patch scheme 4 in terms of Young’s modulus E and Poisson ratio ν.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
208
+ page_content=' To have strong microscale heterogeneity we choose these parameters randomly so that at each microscale grid- point (iid): Eij is log-normal (here varies by factor of about ten);' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
209
+ page_content=' and νij is uniform on [0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
210
+ page_content='25, 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
211
+ page_content='35].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
212
+ page_content=' Figure 3 shows an example Eij.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
213
+ page_content=' Despite such strong heterogeneity, the movie of Figure 1 shows the macroscale dynamics appears relatively simple.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
214
+ page_content=' 2 Equation-free patch scheme Instead of computing the entire beam as seen in Figure 1, the patch scheme computes only in small sparse spatial patches such as Figure 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
215
+ page_content=' In this example case, the patch scheme reduces compute time by a factor ∝ r := (patch size)/(spacing H), which here is just a modest factor of 1/4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
216
+ page_content=' But with greater scale separation and/or in higher spatial dimensions, the scheme often reduces computational time by many orders of magnitude.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
217
+ page_content=' The movie of Figure 4 shows a slow progressive wave of beam bending, together with a not-so-slow compression wave along the beam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
218
+ page_content=' These macroscale predictions are accurate (Section 3) due to the correctness of our simple coupling between patches—even when heterogeneity is strong.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
219
+ page_content=' The patch scheme makes these accurate macroscale predictions even when the macroscale closure is unknown: the scheme does not code a closure.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
220
+ page_content=' Further, ‘the closure’ varies depending upon human assumptions such as choosing averaged models versus cosserat models—the patch scheme makes no such closure assumptions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
221
+ page_content=' The only assumption is that the macroscale quantities of importance vary smoothly between neighbouring patches.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
222
+ page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
223
+ page_content='1 Scheme is non-intrusive functional ‘wrapper’ Consider one of the patches of the 2D beam shown in Figure 4.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
224
+ page_content=' With the given microscale xy-grid (Figure 2), zooming in to the microscale each patch is like that of Figure 5.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
225
+ page_content=' Here each patch extends across the cross-section (y-dimension) of the beam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
226
+ page_content=' Open symbols in Figure 5 are ghost nodes outside the patch and implement given stress-free top/bottom conditions on the beam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
227
+ page_content=' The only addition required by the patch scheme are the edge values (‘squared’ micro-grid nodes in Figure 5) on the left/right of each patch.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
228
+ page_content=' The patch scheme couples patches together by providing the patch-edge values through interpolation across the macroscale between patches (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
229
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
230
+ page_content=', Roberts & Kevrekidis 2007, Roberts et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
231
+ page_content=' 2014, Cao & Roberts 2016).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
232
+ page_content=' Here we interpolate from each of the centre patch values across the beam (i = 4 in Figure 5) of ‘nearby’ Figure 4: movie of a patch scheme simulation of a heterogeneous beam showing the macroscale propagation across the patches of beam bending waves and longitudinal compression waves.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
233
+ page_content=' 0 1 2 3 4 5 6 space x 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
234
+ page_content='2 0 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
235
+ page_content='2 y time = 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
236
+ page_content='00, E in 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
237
+ page_content='35 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
238
+ page_content='2 2 Equation-free patch scheme 5 Figure 5: one example patch of the 2D elastic beam show- ing the microscale staggered grid (Figure 2).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
239
+ page_content=' This is case of nsubpatch = 7 micro-grid in- tervals along the patch, and ny = 4 intervals across the beam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
240
+ page_content=' ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▶ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ▲ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊚ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ⊗ ��� ⊗ ⊗ ▷ ▷ ▷ ▷ ▷ ▷ ⃝ ⃝ ⃝ ⃝ ⃝ ⃝ ▷ ▷ ▷ ▷ ▷ ▷ ⃝ ⃝ ⃝ ⃝ ⃝ ⃝ □ □ □ □ □ □ □ □ □ □ □ □ □ □ i = 1 2 3 4 5 6 7 j =1 2 3 4 patches, to determine the corresponding patch-edge value.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
241
+ page_content=' Here we implement spectral (fft) interpolation between the patches for high accuracy (Section 3).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
242
+ page_content=' The scheme does not presume that any average is appropriate.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
243
+ page_content=' This implementation shows that the patch scheme is non-intrusive (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
244
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
245
+ page_content=', Biezemans et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
246
+ page_content=' 2022): it just ‘wraps around’ any micro-grid code a user trusts.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
247
+ page_content=' Consequently, we provide a toolbox (Maclean et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
248
+ page_content=' 2021) for others to implement the patch scheme around their micro-code.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
249
+ page_content=' 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
250
+ page_content='2 Scheme embeds macroscale dynamics Given the patch scheme does not assume anything about what are ‘correct’ macroscale variables, a crucial question is the following: how can we be assured that the patch scheme captures the macroscale slow dynamics?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
251
+ page_content=' An answer is provided by the Whitney (1936) embedding theorem.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
252
+ page_content=' Roughly, the theorem is that every mD manifold is parametrisable from almost every subspace of more than 2mD.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
253
+ page_content=' Let’s see what this means for us.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
254
+ page_content=' In essence, the patch scheme provides the higher-D subspace in which the slow manifold of the macroscale wave dynamics is embedded.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
255
+ page_content=' For beams in two spatial dimensions, the basic macroscale beam models have, at each cross-section, displacement and velocity of both bending and compression.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
256
+ page_content=' Thus the elastic beam dynamics has a slow manifold that is m = 4D at every cross-section.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
257
+ page_content='1 Alternatively, 2D cosserat beam models add a shear mode to the macroscale model—two more variables—leading to a not-quite-so-slow manifold of m = 6D at every cross-section.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
258
+ page_content=' These physically based models are slow manifolds because they focus on the relatively slow waves of solutions varying slowly in space, and neglect all the faster high-frequency cross-waves.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
259
+ page_content=' In the patch scheme, Figures 1 and 4 show simulations with a cross-section of ny = 7 micro-grid intervals, but let’s discuss the case of just ny = 4 (Figure 5).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
260
+ page_content=' For ny = 4, there are seven microscale nodes across each patch edge.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
261
+ page_content=' Each node has a displacement and velocity, and so leads to a 14D subspace for macroscale communication between patches.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
262
+ page_content=' 1Such statements, invoking a manifold or subspace “at every cross-section”, are in a sense developed by the theory of Roberts (2015).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
263
+ page_content=' That is, in systems of large spatial extent there often are important, spatially global, invariant manifolds of high-D that are effectively decomposable into a union of spatially local manifolds/subspaces of relatively lower dimension—a dimension determined by the spatial cross-section—and that are weakly coupled to neighbouring locales.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
264
+ page_content=' 3 Scheme has proven accuracy 6 Figure 6: multiscale spectrum of eigenvalues λ separates macroscale modes on the right from sub-patch microscale modes on the left.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
265
+ page_content=' The axes are scaled nonlinearly.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
266
+ page_content=' Here the small viscosity is 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
267
+ page_content='001 so the microscale decays, but the macroscale waves are long- lasting.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
268
+ page_content=' 3 1 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
269
+ page_content='3 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
270
+ page_content='1 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
271
+ page_content='03 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
272
+ page_content='01 0 100 30 10 3 1 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
273
+ page_content='3 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
274
+ page_content='10 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
275
+ page_content='1 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
276
+ page_content='3 1 3 10 30 100 ℜλ ℑλ Because 14 > 2 · 6 > 2 · 4 , the Whitney embedding theorem asserts that the patch scheme exchanges enough information to almost surely parametrise both such slow manifolds of the macroscale dynamics.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
277
+ page_content=' The patch scheme does not need to explicitly compute and exchange specific assumed macroscale average quantities.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
278
+ page_content=' 3 Scheme has proven accuracy Section 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
279
+ page_content='2 discusses established theory which generally proves that the patch scheme makes accurate macroscale predictions.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
280
+ page_content=' Such proofs are in stark contrast to the vast machine learning/artificial intelligence developments which prove very few general results: for example, Brenner & Koumoutsakos (2021) comment “.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
281
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
282
+ page_content=' .' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
283
+ page_content=' ml studies, as the lack of rigorous theory does not offer (yet!' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
284
+ page_content=') guarantees of convergence”.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
285
+ page_content=' Before discussing theory, we first report some computational verification of high accuracy.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
286
+ page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
287
+ page_content='1 Computation verifies exactness Here we restricted attention to linear elasticity so we know that the wrapped patch system is fully characterised by the resultant Jacobian matrix.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
288
+ page_content=' We numeri- cally compute the Jacobian matrix of the patch scheme by elementary numerical differentiation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
289
+ page_content=' Because of the macroscale translational invariance of the patch scheme, the macroscale eigenvectors are correctly sinusoidal.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
290
+ page_content=' Hence the only macroscale er- rors occur in the eigenvalues of the Jacobian.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
291
+ page_content=' Figure 6 plots the spectrum of all eigenvalues for one example of random heterogeneity, in the case of five patches for simplicity.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
292
+ page_content=' Observe there are: (on the right) four λ = 0 of rigid beam motion;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
293
+ page_content=' four −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
294
+ page_content='001 ± i 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
295
+ page_content='057 and four −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
296
+ page_content='003 ± i 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
297
+ page_content='111 of compressions waves;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
298
+ page_content=' four −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
299
+ page_content='001 ± i 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
300
+ page_content='061 and four −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
301
+ page_content='004 ± i 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
302
+ page_content='237 of beam bending waves;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
303
+ page_content=' with the above macroscale eigenvalues separated by a spectral gap from the following sub-patch microscale eigenvalues;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
304
+ page_content=' 3 Scheme has proven accuracy 7 Table 1: error in patch scheme’s macroscale eigenvalues λ for various patch size ratios r: the macroscale λs are exact to round- off error—due to patch coupling by spectral interpolation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
305
+ page_content=' macro-eigenvalue r = 1 2 r = 1 4 r = 1 8 −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
306
+ page_content='001 ± i 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
307
+ page_content='061 2e-12 1e-12 2e-13 −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
308
+ page_content='001 ± i 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
309
+ page_content='061 2e-12 4e-12 2e-12 −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
310
+ page_content='004 ± i 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
311
+ page_content='237 1e-12 8e-13 3e-12 −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
312
+ page_content='004 ± i 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
313
+ page_content='237 1e-12 2e-12 3e-12 −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
314
+ page_content='001 ± i 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
315
+ page_content='057 7e-13 4e-13 6e-13 −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
316
+ page_content='001 ± i 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
317
+ page_content='057 6e-13 5e-13 6e-13 −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
318
+ page_content='003 ± i 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
319
+ page_content='111 1e-13 2e-13 2e-13 −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
320
+ page_content='003 ± i 2.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
321
+ page_content='111 4e-13 5e-13 2e-13 Figure 7: multiscale spectrum of eigenvalues λ for the patch scheme in the case of zero viscos- ity.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
322
+ page_content=' The horizontal axis shows that all modes have zero real- part to numerical round-off er- ror.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
323
+ page_content=' That is, in the case of zero viscosity, this patch scheme pre- serves the wave nature of the underlying physics.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
324
+ page_content=' 5e-13 2e-13 1e-130 1e-13 2e-13 5e-13 100 30 10 3 1 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
325
+ page_content='3 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
326
+ page_content='10 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
327
+ page_content='1 0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
328
+ page_content='3 1 3 10 30 100 ℜλ ℑλ (on the left) many ℜλ < −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
329
+ page_content='1 of uninteresting sub-patch micro-scale fast- waves (headed by ten eigenvalues around −0.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
330
+ page_content='14 ± i 9.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
331
+ page_content='29).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
332
+ page_content=' To quantify the accuracy, Table 1 compares eigenvalues obtained from full-domain code, with the above macroscale eigenvalues obtained by the wrapped patch scheme.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
333
+ page_content=' For all patch size ratios and heterogeneities tested, the patch scheme’s macroscale eigenvalues are exact to numerical round-off error.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
334
+ page_content=' Such exactness is due to the spectral interpolation used here.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
335
+ page_content=' If, instead of spectral, local polynomial interpolation of degree p is used to couple the patches, then generally the patch scheme has macroscale errors ∝ Hp where H = inter-patch spacing (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
336
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
337
+ page_content=', Roberts & Kevrekidis 2007, Roberts et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
338
+ page_content=' 2014).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
339
+ page_content=' Undamped waves?' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
340
+ page_content=' With zero viscosity, there are only oscillations in the under- lying physics.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
341
+ page_content=' In such a scenario computational methods are very delicate.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
342
+ page_content=' Here, Figure 7 illustrates that all eigenvalues of the patch scheme have |ℜλ| < 10−12.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
343
+ page_content='2 Hence, even with no viscosity, the patch scheme preserves the oscillatory wave nature of the heterogeneous physics.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
344
+ page_content=' There is a perception that the patch scheme “only works well on problems with an inertial manifold and for systems in which most modes are strongly decaying” (Majda & Grooms 2014, p.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
345
+ page_content='62).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
346
+ page_content=' This verification of accuracy for purely elastic 2In some realisations of the heterogeneity, the sensitive multiplicity four eigenvalue λ = 0 numerically splits into four showing |ℜλ| up to 10−6 due to round-off errors.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
347
+ page_content=' 4 Conclusion 8 beams shows that this perception is false.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
348
+ page_content=' Applications and theory for other wave systems also refute this perception (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
349
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
350
+ page_content=', Cao & Roberts 2016, Bunder et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
351
+ page_content=' 2021, Divahar et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
352
+ page_content=' 2022).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
353
+ page_content=' 3.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
354
+ page_content='2 Mathematical analysis proves consistency Mathematical analysis has proven properties of the patch scheme in general.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
355
+ page_content=' Mostly, the published proofs explicitly address dissipative (nonlinear) systems.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
356
+ page_content=' However, as discussed by Bunder et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
357
+ page_content=' (2021), the patch scheme in space only recasts spatial interactions, so whether the time derivative is ∂/∂t of dissipation or ∂2/∂t2 of waves makes little difference.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
358
+ page_content=' Two complementary types of results have been proven.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
359
+ page_content=' They involve the spacing between patch centres H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
360
+ page_content=' First, Centre Manifold Theory may be applied at finite spacing H by introducing a ‘bookkeeping’ parameter γ to label inter-patch communication (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
361
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
362
+ page_content=', Roberts et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
363
+ page_content=' 2014, §2) to prove the existence of a slow manifold in the patch scheme (including when it is applied to nonlinear systems).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
364
+ page_content=' Then the parameter γ structures inter-patch interactions, and their algebraic expression, to empower theory based at γ = 0, via regular perturbation, to address finite γ such as the case of full coupling γ = 1 (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
365
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
366
+ page_content=', Roberts et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
367
+ page_content=' 2014, Cor.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
368
+ page_content=' 2).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
369
+ page_content=' Second, the patch scheme is consistent with the underlying micro-code as the patch spacing H → 0 (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
370
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
371
+ page_content=', Roberts et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
372
+ page_content=' 2014, Thm.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
373
+ page_content=' 7).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
374
+ page_content=' The consistency is that the macroscale of the patch scheme is the same as the macroscale of the given micro-coded system, to errors O � Hp� when using polynomial interpolation of degree p.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
375
+ page_content=' For example, spectral interpolation corresponds to ‘p = ∞’ so then errors vanish to all orders as in Table 1.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
376
+ page_content=' These results and general proofs were first done for homogeneous systems (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
377
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
378
+ page_content=', Roberts & Kevrekidis 2007, Roberts et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
379
+ page_content=' 2014).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
380
+ page_content=' They were subsequently ex- tended to heterogeneous microscales (Bunder et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
381
+ page_content=' 2017), and recently extended to alternative inter-patch coupling that preserves self-adjointness (Bunder et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
382
+ page_content=' 2021).' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
383
+ page_content=' Interestingly, the extension of the theoretical support to heterogeneous cases invokes the ensemble of all phase-shifts of the heterogeneity.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
384
+ page_content=' The ensemble is spatially homogeneous, so the homogeneous proofs and results apply to establish the heterogeneous results.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
385
+ page_content=' 4 Conclusion As an initial exploration of the patch scheme for homogenisation of heterogeneous elasticity, we considered the prototypical case of a 2D elastic beam.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
386
+ page_content=' The scheme gives a non-intrusive and efficient computational homogenisation of given microscale system via spatially sparse small patches.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
387
+ page_content=' The patch coupling has proven accuracy, controllable error, at finite scale separation.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
388
+ page_content=' The patch scheme makes only one assumption: in the scenarios of interest to a user, there is no significant spatial structures in the mesoscale between the patch spacing H and the microscale resolved in the patches.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
389
+ page_content=' In contrast to most other multiscale methods, there is: no assumed boundary conditions on Representative Volume Elements (variously periodic, Dirichlet, Neumann);' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
390
+ page_content=' no explicitly assuming References 9 slow variables;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
391
+ page_content=' and no presumed necessary variational principle.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
392
+ page_content=' The scheme is entirely physically interpretable: there is no hidden mystic machinations of neural networks (e.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
393
+ page_content='g.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
394
+ page_content=', Brenner & Koumoutsakos 2021) The patch scheme is simple to apply.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
395
+ page_content=' In contrast to other multiscale methods there is: no arbitrary averaging;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
396
+ page_content=' no oversampling regions;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
397
+ page_content=' no buffer regions;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
398
+ page_content=' no action regions;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
399
+ page_content=' no guessed fast/slow variables;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
400
+ page_content=' no epsilons;' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
401
+ page_content=' and no limits.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
402
+ page_content=' As a non-intrusive ‘wrapper’, anyone can start using the patch scheme via a Matlab/ Octave Toolbox (Maclean et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
403
+ page_content=' 2021, Roberts et al.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
404
+ page_content=' 2019–2023) Acknowledgements This research was supported by Australian Research Coun- cil grants DP220103156 and DP200103097.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
405
+ page_content=' References Biezemans, R.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
406
+ page_content=' A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
407
+ page_content=', Le Bris, C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
408
+ page_content=', Legoll, F.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
409
+ page_content=' & Lozinski, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
410
+ page_content=' (2022), Non-intrusive implementation of a wide variety of Multiscale Finite Element Methods, Technical report, http://arxiv.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
411
+ page_content='org/abs/2211.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
412
+ page_content='17024.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
413
+ page_content=' Brenner, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
414
+ page_content=' P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
415
+ page_content=' & Koumoutsakos, P.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
416
+ page_content=' (2021), ‘Editorial: Machine Learning and Physical Review Fluids: An Editorial Perspective’, Physical Review Fluids 6(7), 070001.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
417
+ page_content=' Bunder, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
418
+ page_content=' E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
419
+ page_content=', Kevrekidis, I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
420
+ page_content=' G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
421
+ page_content=' & Roberts, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
422
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
423
+ page_content=' (2021), ‘Equation-free patch scheme for efficient computational homogenisation via self-adjoint coupling’, Numerische Mathematik 149(2), 229–272.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
424
+ page_content=' Bunder, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
425
+ page_content=' E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
426
+ page_content=', Roberts, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
427
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
428
+ page_content=' & Kevrekidis, I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
429
+ page_content=' G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
430
+ page_content=' (2017), ‘Good coupling for the mul- tiscale patch scheme on systems with microscale heterogeneity’, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
431
+ page_content=' Computational Physics 337, 154–174.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
432
+ page_content=' Cao, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
433
+ page_content=' & Roberts, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
434
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
435
+ page_content=' (2016), ‘Multiscale modelling couples patches of nonlinear wave-like simulations’, IMA J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
436
+ page_content=' Applied Maths.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
437
+ page_content=' 81(2), 228–254.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
438
+ page_content=' Divahar, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
439
+ page_content=', Roberts, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
440
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
441
+ page_content=', Mattner, T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
442
+ page_content=' W.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
443
+ page_content=', Bunder, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
444
+ page_content=' E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
445
+ page_content=' & Kevrekidis, I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
446
+ page_content=' G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
447
+ page_content=' (2022), Two novel families of multiscale staggered patch schemes efficiently simulate large-scale, weakly damped, linear waves, Technical report, https: //arxiv.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
448
+ page_content='org/abs/2210.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
449
+ page_content='15823.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
450
+ page_content=' Lucarini, S.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
451
+ page_content=', Upadhyay, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
452
+ page_content=' V.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
453
+ page_content=' & Segurado, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
454
+ page_content=' (2021), ‘FFT based approaches in micromechanics: fundamentals, methods and applications’, Modelling and Simulation in Materials Science and Engineering 30(2), 023002.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
455
+ page_content=' Maclean, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
456
+ page_content=', Bunder, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
457
+ page_content=' E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
458
+ page_content=' & Roberts, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
459
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
460
+ page_content=' (2021), ‘A toolbox of equation-free functions in matlab/octave for efficient system level simulation’, Numerical Algorithms 87, 1729–1748.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
461
+ page_content=' Majda, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
462
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
463
+ page_content=' & Grooms, I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
464
+ page_content=' (2014), ‘New perspectives on superparameterization for geophysical turbulence’, Journal of Computational Physics 271, 60–77.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
465
+ page_content=' Matouˇs, K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
466
+ page_content=', Geers, M.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
467
+ page_content=' G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
468
+ page_content=' D.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
469
+ page_content=', Kouznetsova, V.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
470
+ page_content=' G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
471
+ page_content=' & Gillman, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
472
+ page_content=' (2017), ‘A review of predictive nonlinear theories for multiscale modeling of heterogeneous materials’, Journal of Computational Physics 330, 192–220.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
473
+ page_content=' References 10 Raju, K.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
474
+ page_content=', Tay, T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
475
+ page_content='-E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
476
+ page_content=' & Tan, V.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
477
+ page_content=' B.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
478
+ page_content=' C.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
479
+ page_content=' (2021), ‘A review of the FE2 method for composites’, Multiscale and Multidisciplinary Modeling, Experiments and Design 4, 1–24.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
480
+ page_content=' Roberts, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
481
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
482
+ page_content=' (2015), ‘Macroscale, slowly varying, models emerge from the mi- croscale dynamics in long thin domains’, IMA Journal of Applied Mathematics 80(5), 1492–1518.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
483
+ page_content=' Roberts, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
484
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
485
+ page_content=' & Kevrekidis, I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
486
+ page_content=' G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
487
+ page_content=' (2007), ‘General tooth boundary conditions for equation free modelling’, SIAM J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
488
+ page_content=' Scientific Computing 29(4), 1495–1510.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
489
+ page_content=' Roberts, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
490
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
491
+ page_content=', MacKenzie, T.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
492
+ page_content=' & Bunder, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
493
+ page_content=' (2014), ‘A dynamical systems approach to simulating macroscale spatial dynamics in multiple dimensions’, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
494
+ page_content=' Engineering Mathematics 86(1), 175–207.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
495
+ page_content=' Roberts, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
496
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
497
+ page_content=', Maclean, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
498
+ page_content=' & Bunder, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
499
+ page_content=' E.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
500
+ page_content=' (2019–2023), Equation-free function tool- box for matlab/octave, Technical report, [https://github.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
501
+ page_content='com/uoa1184615/ EquationFreeGit].' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
502
+ page_content=' Samaey, G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
503
+ page_content=', Roberts, A.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
504
+ page_content=' J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
505
+ page_content=' & Kevrekidis, I.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
506
+ page_content=' G.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
507
+ page_content=' (2010), Equation-free computation: an overview of patch dynamics, in J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
508
+ page_content=' Fish, ed.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
509
+ page_content=', ‘Multiscale methods: bridging the scales in science and engineering’, Oxford University Press, chapter 8, pp.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
510
+ page_content=' 216– 246.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
511
+ page_content=' Somnic, J.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
512
+ page_content=' & Jo, B.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
513
+ page_content=' W.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
514
+ page_content=' (2022), ‘Status and challenges in homogenization methods for lattice materials’, Materials 15(2), 605.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
515
+ page_content=' Whitney, H.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
516
+ page_content=' (1936), ‘Differentiable manifolds’, Annals of Mathematics 37(3), 645– 680.' metadata={'source': '/home/zjlab/wf/langchain-ChatGLM/knowledge_base/WtFPT4oBgHgl3EQfrjVl/content/2301.13145v1.pdf'}
Y9E3T4oBgHgl3EQfcgoK/content/tmp_files/2301.04525v1.pdf.txt ADDED
@@ -0,0 +1,1157 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Clustering disease trajectories in contrastive
2
+ feature space for biomarker discovery in
3
+ age-related macular degeneration
4
+ Robbie Holland1(�), Oliver Leingang2, Christopher Holmes3, Philipp Anders4,
5
+ Johannes C. Paetzold1, Rebecca Kaye7, Sophie Riedl2, Hrvoje Bogunović2,
6
+ Ursula Schmidt-Erfurth2, Lars Fritsche6, Hendrik P. N. Scholl4,5, Sobha
7
+ Sivaprasad3, Andrew J. Lotery7, Daniel Rueckert1,8, Martin J. Menten1,8
8
+ 1 BioMedIA, Imperial College London, London, United Kingdom
9
10
+ 2 Laboratory for Ophthalmic Image Analysis, Medical University of Vienna
11
+ 3 Moorfields National Institute for Health and Care Biomedical Research Centre,
12
+ Moorfields Eye Hospital, London, United Kingdom
13
+ 4 Institute of Molecular and Clinical Ophthalmology Basel
14
+ 5 Department of Ophthalmology, Universitat Basel, Basel, Switzerland
15
+ 6 Department of Biostatistics, University of Michigan, Ann Arbor, United States
16
+ 7 Clinical and Experimental Sciences, Faculty of Medicine, University of
17
+ Southampton, Southampton, United Kingdom
18
+ 8 Institute for AI and Informatics in Medicine, Technical University of Munich,
19
+ Munich, Germany
20
+ Abstract. Age-related macular degeneration (AMD) is the leading cause
21
+ of blindness in the elderly. Despite this, the exact dynamics of disease
22
+ progression are poorly understood. There is a clear need for imaging
23
+ biomarkers in retinal optical coherence tomography (OCT) that aid the
24
+ diagnosis, prognosis and management of AMD. However, current grad-
25
+ ing systems, which coarsely group disease stage into broad categories
26
+ describing early and intermediate AMD, have very limited prognostic
27
+ value for the conversion to late AMD. In this paper, we are the first to
28
+ analyse disease progression as clustered trajectories in a self-supervised
29
+ feature space. Our method first pretrains an encoder with contrastive
30
+ learning to project images from longitudinal time series to points in
31
+ feature space. This enables the creation of disease trajectories, which
32
+ are then denoised, partitioned and grouped into clusters. These clusters,
33
+ found in two datasets containing time series of 7,912 patients imaged
34
+ over eight years, were correlated with known OCT biomarkers. This re-
35
+ inforced efforts by four expert ophthalmologists to investigate clusters,
36
+ during a clinical comparison and interpretation task, as candidates for
37
+ time-dependent biomarkers that describe progression of AMD.
38
+ Keywords: Contrastive learning · Trajectory clustering · Disease pro-
39
+ gression · Retina · OCT · Biomarker discovery
40
+ arXiv:2301.04525v1 [eess.IV] 11 Jan 2023
41
+
42
+ 2
43
+ Holland et al.
44
+ 1
45
+ Introduction
46
+ AMD is the leading cause of blindness in the elderly, affecting nearly 200 mil-
47
+ lion people worldwide [22]. Patients with early stages of the disease exhibit
48
+ few symptoms until suddenly converting to the late stage, at which point their
49
+ sharp central vision rapidly deteriorates [14]. AMD patients are commonly im-
50
+ aged with optical coherence tomography (OCT), which provides low-cost and
51
+ high-resolution retinal images that depict fine physiological details. Several OCT
52
+ biomarkers have been tentatively linked to the pathogenesis of AMD, such as reti-
53
+ nal layer thickness, photoreceptor atrophy and the presence of drusen, which are
54
+ lipidic deposits that build up inside the retina [16]. However, the exact relation
55
+ of these biomarkers to the progression from early to late stages of AMD remains
56
+ unclear. Current grading systems only coarsely group patients into broad cate-
57
+ gories for early, intermediate and late AMD and have limited prognostic value
58
+ [8]. There is an unmet need for biomarkers that describe and predict the pro-
59
+ gression of AMD.
60
+ Large studies that image populations of AMD patients over time are a power-
61
+ ful resource to discover novel biomarkers for disease progression. If candidate
62
+ biomarkers are already identified a priori, they can be extracted from images
63
+ and mapped against disease progression [16,19,3]. However, this approach is not
64
+ feasible if the biomarkers are partially or fully unknown. Self-supervised learn-
65
+ ing is the most promising method to automatically discover new biomarkers by
66
+ clustering or finding anomalous OCT images in lower-dimensional representation
67
+ space [18,21,17]. However, by grouping single scans acquired at single points in
68
+ time, these biomarkers are by definition static and cannot capture concepts such
69
+ as the speed of disease progression or transitions between multiple states of the
70
+ disease. Clustering whole time series of images is also problematic, as patients
71
+ enter and leave the study at different points in their overall progression. There-
72
+ fore, in order to discover biomarkers that describe population-level patterns in
73
+ disease progression, it is necessary to analyse and compare portions of patient
74
+ time series.
75
+ Our contribution In this work, we develop a strategy to automatically discover
76
+ biomarkers that capture disease progression in time series of images. This is
77
+ illustrated in Figure 1. Our method represents the time series of images as a
78
+ trajectory in a self-supervised latent feature space built by contrastive learning.
79
+ This representation allows a new partitioning of time series of OCT images
80
+ into consecutive subsequences (sub-trajectories) that exhibit distinct units of
81
+ disease progression. By clustering these sub-trajectories, we are now able to
82
+ detect patterns of disease progression that are common among the population
83
+ of patients.
84
+ Experimentally, we test our method on a large cohort from two, longitudinal
85
+ retinal OCT datasets totalling 160,558 images from 7,912 patients. In doing so we
86
+ categorise 3,218 total years of disease progression with time-dependent clusters.
87
+ We then correlate our clusters with the known set of biomarkers which reinforces
88
+
89
+ Clustering disease trajectories for biomarker discovery in AMD
90
+ 3
91
+ Fig. 1: We analyse disease progression as clustered trajectories in feature space.
92
+ We first train a feature extractor with contrastive learning and then form tra-
93
+ jectories by projecting time series of images to feature space. These are then
94
+ partitioned into a set of sub-trajectories which are clustered. Finally, clusters are
95
+ related to known disease stages before being interpreted as candidate biomarkers
96
+ for disease state and progression. PCA space is coloured by visual acuity.
97
+ their potential to contain potentially new, time-dependent biomarkers. Finally,
98
+ we closed the loop on automated biomarker discovery by working directly with
99
+ four ophthalmologists to interpret these clusters in a clinical comparison task.
100
+
101
+ 1. Contrastive pretraining
102
+ Contrastive
103
+ PCA feature
104
+ Retinal OCT datasets (170k scans)
105
+ loss
106
+ space
107
+ ResNet-50
108
+ CNN
109
+ 00
110
+ 0
111
+ 00
112
+ 0
113
+ Copy pretrained
114
+ feature extractor
115
+ 0
116
+ 0
117
+ 0
118
+ PCA reduction
119
+ 00
120
+ ResNet-50
121
+ 00
122
+ CNN
123
+ 00
124
+ 00
125
+ 2. Patient trajectories in feature space
126
+ Resample
127
+ Longitudinal information
128
+ with temporal
129
+ denoising kernel
130
+ Eye
131
+ Scan times
132
+ 04/2017
133
+ 08/2017
134
+ 11/2017
135
+ 2
136
+ 07/2012
137
+ 12/2012
138
+ 04/2013
139
+ 3
140
+ 06/2019
141
+ 09/2020
142
+ 10/2020
143
+ 01/2015
144
+ 03/2016
145
+ 04/2016
146
+ 5
147
+ 08/2017
148
+ 09/2017
149
+ 02/2018
150
+ 6
151
+ 09/2012
152
+ 03/2013
153
+ 07/2013
154
+ 3. Clustering and proposing biomarkers
155
+ Cluster 1 vs. Cluster 2
156
+ Series A
157
+ K-means
158
+ Query series
159
+ Count biomarkers
160
+ clustering
161
+ Known biomarkers
162
+ Series B
163
+ per cluster
164
+ Healthy Drusen cRORA
165
+ CNV
166
+ 0
167
+ Clusters
168
+ Experts compare clusterst with
169
+ similar known biomarkers
170
+ 2
171
+ but different latent features
172
+ tclusters are candidates for
173
+ time-dependent biomarkers4
174
+ Holland et al.
175
+ 2
176
+ Related work
177
+ 2.1
178
+ Current AMD grading systems
179
+ Ophthalmologists’ current understanding of progression from early to late AMD
180
+ largely involves drusen. Drusen can grow until suddenly regressing and disap-
181
+ pearing, which often precedes the onset of late AMD [16]. While there have
182
+ been attempts to group drusen based on their morphology [11], current grading
183
+ systems stratify early and intermediate stages solely by drusen size [2,12,7,8].
184
+ Late AMD is classified as either choroidal neovascularisation (CNV), identified
185
+ by fluid under the retina, or geographic atrophy by progressive loss of photore-
186
+ ceptors and retinal thinning. The degree of atrophy can be staged using cRORA
187
+ (complete retinal pigment epithelium and outer retinal atrophy), which measures
188
+ the width of focal atrophy in OCT [15]. So far grading systems offer limited di-
189
+ agnostic value and little to no prognostic value.
190
+ 2.2
191
+ Tracking evolution of known biomarkers
192
+ Few research efforts have aimed at quantifying and tracking known AMD biomark-
193
+ ers over time, such as reticular pseudodrusen [19] and drusen volume [16]. More
194
+ work has explored Alzheimer’s disease (AD), which offers a greater array of
195
+ quantitative imaging biomarkers, such as levels of tau protein and hippocampal
196
+ volume. Young et al. [23] fit an event-based model that rediscovers the order
197
+ in which these biomarkers become anomalous as AD progresses. Vogel et al.
198
+ [20] find four distinct spatiotemporal trajectories for tau pathology in the brain.
199
+ However, mapping biomarkers that evolve during disease progression requires
200
+ prior annotation of entire time series. Thus, these biomarkers must be known or
201
+ at least suspected a priori.
202
+ 2.3
203
+ Automated discovery of unknown biomarkers
204
+ In order to discover new biomarkers, efforts to find them have turned to au-
205
+ tomated biomarker discovery. Imaging biomarkers are proposed by analysing
206
+ anomalous scans [17], clusters of scans [25], or a combination of these [18] in
207
+ feature space. To build these, neural networks are trained with supervised or
208
+ unsupervised proxy tasks. These tasks include image reconstruction [21], seg-
209
+ mentation [25] and generative adversarial networks [17]. However, networks are
210
+ prone to overfit on their specific task and lose semantic information regarding
211
+ the disease. Contrastive learning has recently advanced the state-of-the-art in
212
+ training generalisable and unbiased feature extractors. Chen et al. popularised
213
+ this paradigm with SimCLR [4], which was later improved on by Grill et al. [9] in
214
+ Bootstrap Your Own Latent (BYOL). Contrastive methods encode invariance to
215
+ a set of transformations typically uncorrelated with disease features, including
216
+ rotation, translation and global shifts in image brightness and contrast. Zhao et
217
+ al. leverage contrastive feature spaces to identify high-risk clusters of CT image
218
+ patches [24].
219
+
220
+ Clustering disease trajectories for biomarker discovery in AMD
221
+ 5
222
+ However, all biomarkers discovered by the aforementioned methods work by
223
+ grouping single images acquired at single points in time, and in doing so neglect
224
+ temporal relationships between images of the same subject. One work that tack-
225
+ les this challenge, and the most related to ours, categorises the time-dependent
226
+ response of cancer cells to drugs, measured by the changing distance in con-
227
+ trastive feature space from healthy controls [5].
228
+ 2.4
229
+ Trajectory clustering
230
+ The separate field of trajectory clustering is largely focussed on discovering move-
231
+ ment patterns taken by cars, animals and hurricanes [6,1,13]. Lee et al., in their
232
+ state-of-the-art work TRACLUS [13], assume these trajectories are composed of
233
+ consecutive series of common sub-trajectories. For example, different car jour-
234
+ neys may at some point travel down the same road. Using this principle, they
235
+ develop a partition-and-group framework to cluster segments that are repeated
236
+ across multiple trajectories. Similarly, we assume that disease progression can
237
+ be divided into multiple, common disease pathways. Firstly, this allows us to
238
+ work seamlessly with temporally unaligned scanning series. Secondly, we can
239
+ automatically discover novel disease pathways by interpreting sub-trajectories
240
+ that are shared by multiple AMD patients.
241
+ 3
242
+ Materials and Methods
243
+ 3.1
244
+ Self-supervised feature space using contrastive learning
245
+ We adapt BYOL [9] with update coefficient τ = 0.9995 for contrastive pretrain-
246
+ ing of a ResNet50 (4x) model. As several of the contrastive transformations
247
+ designed for natural images are inapplicable to medical images, we use the set
248
+ tailored for retinal OCT images by Holland et al. [10]. Models were trained on
249
+ the entire dataset for 120,000 steps with momentum 0.9 and a learning rate of
250
+ 5 · 10−4 using the Adam optimiser.
251
+ After pretraining, we first remove any final linear layers before projecting all
252
+ labelled images to the feature space of 2048 dimensions. We then reduce the di-
253
+ mension further using principle component analysis (PCA) with D components.
254
+ Using PCA allows us to interpolate the feature space and results in fewer di-
255
+ mensions, which is advantageous for clustering. To validate that the contrastive
256
+ feature space encodes meaningful information for AMD biomarkers, and find
257
+ the optimal dimension D ∈ {2, 10, 20, 50}, we perform multi-class classification
258
+ of known biomarkers (including healthy controls). Firstly, we split the dataset
259
+ into train and test partitions using 85% and 15% of the data, respectively, en-
260
+ suring that all scans from each patient belong to the same set. Then to perform
261
+ the classification, we fit a class-balanced support vector machine (SVM) on the
262
+ training set and report performance on the test set.
263
+
264
+ 6
265
+ Holland et al.
266
+ Fig. 2: Longitudinal scans of a single eye, imaged over four years, projected to
267
+ PCA space. PCA space is plotted as a hexmap coloured by the local average
268
+ in visual acuity, where higher values indicate poorer quality of vision. Each
269
+ row depicts two principle components up to 20. The rightmost columns show
270
+ resampled trajectories. Using smaller T captures short-term variation in disease
271
+ progression but cannot model long-term changes. MDL aims to optimise this
272
+ tradeoff by partitioning trajectories only at points of inflection.
273
+ 3.2
274
+ Extracting and clustering common sub-trajectories
275
+ For each eye, we first form piecewise-linear trajectories by linking points in PCA
276
+ space that were derived from consecutively acquired OCT images (see left column
277
+ in Figure 2). We then assume, in analogy to TRACLUS [13], that trajectories
278
+ encoding disease progression can be partitioned into sub-trajectories that are
279
+ common among multiple patients. We compare two methods to achieve this,
280
+ shown in the right-most columns in Figure 2. The first method resamples time-
281
+ points at regular intervals of T years. For each resampled time t, we find the
282
+ corresponding point in feature space by taking a weighted average of all points
283
+ in the trajectory. The weights are calculated by convolution of a Gaussian kernel
284
+ N(t, σT ) with the acquisition dates of the entire scanning series. We then define
285
+
286
+ Resampled trajectories
287
+ (T = 0.5 years)
288
+ T=
289
+ T=1.0
290
+ T= 2.0
291
+ T= 0.5
292
+ Original
293
+ Cumulative
294
+ MDL Partition
295
+ (years)
296
+ trajectory
297
+ (years)
298
+ (years)
299
+ variance
300
+ Dims 1-2
301
+ PCA #2
302
+ PCA #2
303
+ +18%
304
+ CA#2
305
+ 18%
306
+ PCA #1
307
+ PCA #1
308
+ PCA #1
309
+ PCA #1
310
+ PCA #1
311
+ 1
312
+ Dims 3-4
313
+ PCA #4
314
+ PCA #4
315
+ PCA#4
316
+ +13%
317
+ 7#
318
+ CA
319
+ 31%
320
+ PCA #3
321
+ PCA #3
322
+ PCA #3
323
+ PCA #3
324
+ PCA #3
325
+ Dims 5-6
326
+ PCA #6
327
+ PCA #6
328
+ +10%
329
+ PCA #6
330
+ PCA #6
331
+ CA #6
332
+ 41%
333
+ PCA #5
334
+ PCA #5
335
+ PCA #5
336
+ PCA #5
337
+ PCA #5
338
+ Dims 7-20
339
+ PCA #20
340
+ PCA #20
341
+ PCA #20
342
+ PCA #20
343
+ #20
344
+ +31%
345
+ PCA
346
+ 72%
347
+ PCA #19
348
+ PCA #19
349
+ PCA #19
350
+ PCA #19
351
+ PCA #19
352
+ Colour map for
353
+ visual acuity
354
+ 0.4
355
+ 0.6
356
+ 0.8
357
+ 1.2
358
+ 1.4
359
+ 0.2
360
+ 1.0
361
+ 1.6
362
+ 1.8Clustering disease trajectories for biomarker discovery in AMD
363
+ 7
364
+ sub-trajectories as vectors between consecutive points that are less than or equal
365
+ to T years apart.
366
+ The second method aims to describe trajectories using the fewest points. It be-
367
+ gins by resampling trajectories using the first method (with intervals of T = 0.5
368
+ years). Then, using the minimum description length (MDL) principle, a mini-
369
+ mal subset of points are chosen that best preserve changes in disease over time.
370
+ To achieve this we find the trajectory H, containing a subset of the points in
371
+ the resampled trajectory O, which minimises the following objective (using the
372
+ greedy solution from [13])
373
+ L(O|H) =
374
+
375
+ p∈O
376
+ d⊥(p, H) − λ|H|
377
+ where d⊥(p, H) is the perpendicular distance from a point p to the piece-wise
378
+ linear trajectory H, and |H| is the number of points in H. The coefficient λ is
379
+ proportional to the total standard deviation in the feature space explained by
380
+ D PCA dimensions.
381
+ Clustering sub-trajectories To cluster common sub-trajectories we require
382
+ a distance function that measures the similarity between vectors. Given two
383
+ sub-trajectories U = (ustart, uend) and V = (vstart, vend) their distance is simply
384
+ d(U, V ) = ∥ustart − vstart∥2 + ∥uend − vend∥2
385
+ Finally, using d we separate sub-trajectories into K clusters using k-means clus-
386
+ tering.
387
+ 3.3
388
+ Finding optimal hyperparameters using the set of known
389
+ biomarkers
390
+ We now search for optimal values for the sub-trajectory time interval T ∈
391
+ {0.5, 1.0, 2.0} years, resampling kernel width σT ∈ {0.25, 0.5, 1.0} years and the
392
+ number of clusters K ∈ {5, 10, 15, 30} using five random seeds for k-means clus-
393
+ tering. To quantitatively compare configurations, we use the conditional entropy
394
+ H(B|C) = H(B, C) − H(C) as a scalar measure of how well the clusters C redis-
395
+ cover the known biomarkers B detailed in section 3.5. This is calculated directly
396
+ from their joint distribution P(B, C), which is found by counting all biomarkers
397
+ recorded within sub-trajectories of each cluster. To ensure the equal contribution
398
+ of all biomarkers, we reweight their marginal distribution P(B) to be uniformly
399
+ distributed. As the number of clusters K increases even randomly permuted as-
400
+ signments p(C) will result in reduced values of H(B|p(C)). We address this by
401
+ using the adjusted reduction in conditional entropy, H′ (using r = 5 random
402
+ trials), where higher values correspond to better rediscovery of B
403
+ H′(B|C) = 1
404
+ r
405
+ r
406
+
407
+ i
408
+ H(B|p(C)) − H(B|C)
409
+ As our ultimate goal is to detect biomarkers beyond the known set, we use H′
410
+ only as an indication for the most suitable configuration.
411
+
412
+ 8
413
+ Holland et al.
414
+ 3.4
415
+ Proposing clusters as candidate biomarkers
416
+ We first relabel the clusters C in order of median visual acuity, so that higher
417
+ cluster numbers indicate poorer quality of vision. In order to see which disease
418
+ stages each cluster describes, we calculate the conditional probability P(B|C) =
419
+ P(B, C)/P(C). Then, to discover new biomarkers beyond the existing set, we
420
+ compare clusters that differ maximally in feature space but minimally in the set
421
+ of known biomarkers. To find these, we explore eleven pairs of distinct clusters
422
+ Ci and Cj with a high degree of cosine similarity P(B|Ci) · P(B|Cj).
423
+ Interpreting candidate biomarkers To examine clusters for candidate biomark-
424
+ ers, we collaborate with four expert ophthalmologists. For each pair of clusters,
425
+ we generate four random ‘A or B’ single-choice questions. Clinicians are shown
426
+ one query series in image space from cluster Ci and two further series denoted
427
+ A and B, one from Ci and the other from Cj. For each question, clinicians are
428
+ tasked with determining which of A or B belongs to the same cluster as the
429
+ query. After completing four questions they are prompted to explain on what
430
+ basis they matched A or B to the query. This format allows us to both assess
431
+ whether the clusters are visually distinguishable by experts and, if so, potentially
432
+ extract descriptions of novel biomarkers.
433
+ 3.5
434
+ OCT datasets
435
+ We apply our method to two independent retinal OCT datasets called Dataset
436
+ A and Dataset B. We developed our method on Dataset A but run experiments
437
+ on both datasets. In both, images were acquired using Topcon 3D OCT devices
438
+ (Topcon Corporation, Tokyo, Japan). After strict quality control, Dataset A
439
+ consists of 46,496 scans of 6,236 eyes from 3,456 patients. Eyes were scanned 7.7
440
+ times over 1.9 years on average at irregular time intervals. The second dataset,
441
+ Dataset B, is larger, containing 114,062 scans of 7,253 eyes from 3,819 patients.
442
+ Eyes were scanned 16.6 times over 3.5 years on average. Of each 3D OCT vol-
443
+ ume, we extracted the transverse 2D slice centred at the fovea and resampled
444
+ to 208×256 pixels with a pixel size of 7.0×23.4 µm2, half the median resolution.
445
+ Each scan is labelled with visual acuity, a functional measure assessing the qual-
446
+ ity of vision measured in LogMAR.
447
+ To record conversions to a comprehensive set of known biomarkers B, we used
448
+ established AMD grading protocols described in section 2.1. Early AMD is char-
449
+ acterised by small drusen between 63-125µm in diameter. We also recorded CNV,
450
+ cRORA (of at least 250µm but smaller than 1000 µm) and cRORA (of at least
451
+ 1000 µm) [15]. Overall, 861 conversion times t0 were recorded, and any sub-
452
+ sequent visits at times t+ before the next conversion were automatically as-
453
+ signed with a separate label. Visits prior to any biomarker were labelled as
454
+ ‘NoBiomarker’. Finally, in each dataset, an additional 150 healthy images that
455
+ exhibit no pathology were recorded. Combining these, the known set of biomark-
456
+ ers B includes 10 biomarkers and labels.
457
+
458
+ Clustering disease trajectories for biomarker discovery in AMD
459
+ 9
460
+ 4
461
+ Results
462
+ Fig. 3: Confusion matrices for multi-class classification of known biomarkers and
463
+ healthy images using D numbers of PCA dimensions. In general, performance
464
+ increases with the number of dimensions D. We find that using D = 20 PCA
465
+ dimensions achieves linear separability between known biomarkers.
466
+ 4.1
467
+ Finding the optimal set of hyperparameters using the known
468
+ biomarkers
469
+ In both Dataset A and Dataset B 20 principal dimensions achieves linear sepa-
470
+ rability between known biomarkers (see Figure 3). Both the healthy stage and
471
+ the only extractable early biomarker, drusen, were found to be highly linearly
472
+ separable. Thus, we use D = 20 for the remainder of our analysis.
473
+ We find that K = 15 clusters of sub-trajectories best explained the set of known
474
+ biomarkers B (Figure 4a) as measured by higher values of H′(B|C). Greater
475
+ values of T, in addition to MDL partitioning, result in decreased H′(B|C).
476
+ We suspect that this is because the known set of biomarkers describe disease
477
+ states rather than state transitions, so they are better captured by shorter sub-
478
+ trajectories. In order to select a configuration that finds clusters evidencing pro-
479
+ gression, we choose T = 1.0, σT = 0.5 and K = 15 for the remainder of our
480
+ analysis.
481
+ 4.2
482
+ Sub-trajectory clusters go beyond the known set of biomarkers
483
+ We find that our clusters encode the set of known biomarkers in Dataset A. As
484
+ seen in Figure 5, clusters effectively separate healthy, early stage and late-stage
485
+
486
+ Dataset A
487
+ 10D
488
+ 20D
489
+ 50D
490
+ 2D
491
+ 0.28
492
+ 0.05
493
+ 0.07
494
+ 0.94
495
+ 0.04
496
+ 0.06
497
+ 0.01
498
+ Healthy
499
+ 0.57
500
+ 0.10
501
+ 0.92
502
+ 0.01
503
+ 0.00
504
+ 0.01
505
+ 0.00
506
+ 0.00
507
+ 0.92
508
+ 0.01
509
+ 0.01
510
+ 0.00
511
+ 0.00
512
+ 0.18
513
+ 0.01
514
+ 0.01
515
+ 0.12
516
+ 0.72
517
+ 0.12
518
+ Drusen
519
+ 0.42
520
+ 0.37
521
+ 0.01
522
+ 0.01
523
+ 0.87
524
+ 0.06
525
+ 0.01
526
+ 0.04
527
+ 0.90
528
+ 0.04
529
+ 0.00
530
+ 0.04
531
+ 0.00
532
+ 0.04
533
+ 0.12
534
+ 0.36
535
+ 0.16
536
+ 0.30
537
+ 0.01
538
+ 0.05
539
+ 0.12
540
+ 0.23
541
+ 0.26
542
+ 0.20
543
+ 0.29
544
+ 0.28
545
+ 0.24
546
+ 0.33
547
+ 0.24
548
+ 0.07
549
+ 0.19
550
+ 0.23
551
+ cRORA (250 μm)
552
+ 0.26
553
+ 0.06
554
+ 0.02
555
+ 0.62
556
+ 0.02
557
+ cRORA (1000 μm)
558
+ 0.04
559
+ 0.16
560
+ 0.16
561
+ 0.62
562
+ 0.02
563
+ 0.07
564
+ 0.22
565
+ 0.01
566
+ 0.29
567
+ 0.02
568
+ 0.63
569
+ 0.07
570
+ 0.62
571
+ 0.06
572
+ 0.05
573
+ 0.06
574
+ 0.21
575
+ 0.19
576
+ 0.42
577
+ CNV
578
+ 0.05
579
+ 0.09
580
+ 0.19
581
+ 0.62
582
+ 0.05
583
+ 0.05
584
+ 0.08
585
+ 0.25
586
+ 0.03
587
+ 0.09
588
+ 0.20
589
+ 0.52
590
+ 0.08
591
+ 0.14
592
+ 0.57
593
+ 0.20
594
+ 0.16
595
+ 0.02Dataset B
596
+ 0.49
597
+ 0.24
598
+ 0.00
599
+ 0.92
600
+ 0.08
601
+ 0.00
602
+ 0.00
603
+ 0.84
604
+ 0.03
605
+ 0.00
606
+ 0.97
607
+ 0.00
608
+ 0.00
609
+ 0.00
610
+ 0.03
611
+ Healthy
612
+ 0.16
613
+ 0.11
614
+ 0.00
615
+ 0.08
616
+ 0.05
617
+ 0.08
618
+ Drusen
619
+ 0.36
620
+ 0.39
621
+ 0.07
622
+ 0.30
623
+ 0.04
624
+ 0.05
625
+ 0.14
626
+ 0.06
627
+ 0.14
628
+ 0.62
629
+ 0.19
630
+ 0.15
631
+ 0.16
632
+ 0.61
633
+ 0.11
634
+ 0.02
635
+ 0.16
636
+ 0.34
637
+ 0.00
638
+ 0.42
639
+ 0.13
640
+ 0.00
641
+ 0.18
642
+ 0.00
643
+ cRORA (250 μm)
644
+ 0.26
645
+ 0.21
646
+ 0.11
647
+ 0.32
648
+ 0.39
649
+ 0.16
650
+ 0.10
651
+ 0.00
652
+ 0.45
653
+ 0.31
654
+ 0.06
655
+ 0.39
656
+ 0.35
657
+ 0.13
658
+ 0.03
659
+ 0.04
660
+ 0.33
661
+ 0.08
662
+ cRORA (1000 μm)
663
+ 0.33
664
+ 0.21
665
+ 0.08
666
+ 0.00
667
+ 0.00
668
+ 0.04
669
+ 0.29
670
+ 0.33
671
+ 0.33
672
+ 0.38
673
+ 0.00
674
+ 0.00
675
+ 0.71
676
+ 0.12
677
+ 0.17
678
+ 0.00
679
+ 0.54
680
+ 0.61
681
+ 0.40
682
+ 0.29
683
+ 0.23
684
+ 0.08
685
+ 0.00
686
+ 0.02
687
+ 0.13
688
+ 0.13
689
+ 0.19
690
+ 0.53
691
+ CNV
692
+ 0.00
693
+ 0.17
694
+ 0.14
695
+ 0.12
696
+ 0.57
697
+ 0.00
698
+ 0.17
699
+ 0.10
700
+ 0.11
701
+ ANO
702
+ CRORA
703
+ CRORA (
704
+ (250
705
+ A (250 μm)10
706
+ Holland et al.
707
+ (a) Adjusted conditional entropy H′(B|C) of known biomarkers B given subtrajectory
708
+ clusters C against hyperparameters K (left), T (center), σT (right)
709
+ Fig. 4: Results of the hyperparameter search, measured by H′(B|C) which is a
710
+ scalar measure of how well the clusters C rediscover existing biomarkers B. As
711
+ we aim to discover biomarkers beyond the known set, we also consider the level
712
+ of disease progression captured by our clusters when choosing our configuration.
713
+ Fig. 5: Conditional probabilities P(B|C) of the known set of biomarkers B given
714
+ cluster assignments C by our method. Cluster pairs highlighted in blue were,
715
+ due to their similarity under the known set of biomarkers B, chosen for further
716
+ analysis by clinicians. One cluster pair, highlighted in pink, was used as a trial
717
+ task and validation experiment.
718
+
719
+ Entropy
720
+ 0.6
721
+ 0.5
722
+ Adjusted reduction in Conditional
723
+ 0.4
724
+ 0.3
725
+ 0.2
726
+ 0.1
727
+ 0.0
728
+ Variants
729
+ Dataset A
730
+ -0.1
731
+ Dataset B
732
+ 5
733
+ 10
734
+ 15
735
+ 30
736
+ Number of clusters (K)Entropy
737
+ 0.6
738
+ 0.5
739
+ 0.4
740
+ 0.3
741
+ 0.2
742
+ 0.1
743
+ 0.0
744
+ Variants
745
+ Dataset A
746
+ -0.1
747
+ Dataset B
748
+ 0.25
749
+ 0.5
750
+ 1.0
751
+ OTEntropy
752
+ 0.6
753
+ 0.5
754
+ Adjusted reduction in Conditional
755
+ 0.4
756
+
757
+ 0.3
758
+ 0.2
759
+ 0.1
760
+ 0.0
761
+ Variants
762
+ Dataset A
763
+ -0.1
764
+ Dataset B
765
+ T= 0.5
766
+ T= 1.0
767
+ T= 2.0
768
+ T= MDL
769
+ partition
770
+ Progression vector sampling strategyDataset A
771
+ Cosine similarities over clusters pairs
772
+ Conditional probability of known biomarkers
773
+ P(BICi) · P(BC))
774
+ given cluster assignments: P(B|C)
775
+ 0.94
776
+ D.55 0.74 0.58 0.75 C
777
+ 0.56 0.63 0.43 0.82
778
+ 2 0.400.15 0.05 0.07
779
+ 0.19
780
+ 0.16
781
+ 0.24
782
+ 0.20
783
+ 0.00
784
+ 0.00
785
+ 0.00
786
+ 0.00
787
+ 0.09
788
+ 0.05
789
+ 0.25
790
+ 1.00
791
+ 0.21
792
+ 0.06
793
+ 0.04
794
+ 0.00
795
+ 0.00
796
+ 0.09
797
+ 0.04
798
+ 0.94 1.00
799
+ 0.34
800
+ 0.84 0.67 0.77 0.67 0.72 0.44 0.71
801
+ 0.49 0.17 0.07 0.08 0.18
802
+ 2
803
+ 0.07
804
+ 0.19
805
+ 0.31
806
+ 2
807
+ 0.70
808
+ 0.12
809
+ 0.07
810
+ 0.07
811
+ 0.00
812
+ 0.00
813
+ 0.00
814
+ 0.00
815
+ 0.02
816
+ 0.02
817
+ 0.55 0.34
818
+ 1.00 0.16 0.14
819
+ 0.46 0.16 0.17 0.37
820
+ 0.74
821
+ 0.08 0.17 0.01 0.06 0.03
822
+ 3
823
+ 3
824
+ 0.00
825
+ 0.15
826
+ 0.22
827
+ 0.04
828
+ 0.07
829
+ 0.74 0.84
830
+ 40.16
831
+ 1.00
832
+ 0.91
833
+ 0.86 0.89 0.96
834
+ 4
835
+ 0.14
836
+ 0.17
837
+ 0.11
838
+ 0.00
839
+ 0.10
840
+ 0.60 0.58 0.73 0.33 0.21 0.23 0.27
841
+ 4
842
+ 5
843
+ 0.00
844
+ 0.17
845
+ 0.07
846
+ 0.12
847
+ 0.16
848
+ 0.21
849
+ 0.07
850
+ 0.03
851
+ 0.11
852
+ 0.06
853
+ 5
854
+ 0.58 0.67
855
+ 0.14 0.91 1.00 0.86 0.97 0.94
856
+ 0.65 0.47 0.82 0.510.34 0.36 0.32
857
+ 9
858
+ 0.11
859
+ 0.09
860
+ 0.10
861
+ 0.11
862
+ 0.08
863
+ 0.04
864
+ 0.12
865
+ 0.75 0.77 0.46
866
+ 6 0.86 0.86 1.00
867
+ 0.91
868
+ 0.89 0.82 0.82 0.82 0.60 0.44 0.46 0.52
869
+ 0.12
870
+ 0.11
871
+ 0.12
872
+ 6
873
+ 6 0.89 0.97 0.91 1.00
874
+ 0.02
875
+ 0.17
876
+ 0.19
877
+ 0.11
878
+ 0.04
879
+ 0.08
880
+ 0.56 0.67(
881
+ 0.92
882
+ D.710.51
883
+ 0.83 0.61 0.45 0.47 0.36
884
+ 7
885
+ 0.11
886
+ 0.12
887
+ 0.09
888
+ 0.07
889
+ 7
890
+ 0.16
891
+ 0.01
892
+ 0.20
893
+ 0.03
894
+ 0.63 0.72 0.17 0.96 0.940.89 0.92 1.00
895
+ 0.76 0.57 0.87 0.44 0.29 0.31 0.43
896
+ 8
897
+ 0.13
898
+ 0.08
899
+ 0.16
900
+ 0.13
901
+ 0.03
902
+ 0.12
903
+ 0.11
904
+ 8
905
+ 0.13
906
+ 0.03
907
+ 0.07
908
+ 0.03
909
+ 0.16
910
+ 0.14
911
+ 0.00
912
+ 0.00
913
+ 0.16
914
+ 0.29
915
+ 0.43 0.44 0.37 0.60 0.65 0.82 0.71 0.76 1.00 0.66 0.90 0.42 0.26 0.27 0.66
916
+ 9
917
+ 9
918
+ 0.82 0.71 0.74 0.58 0.47 0.82 0.51 0.57 0.66
919
+ 0.25
920
+ 0.07
921
+ 0.14
922
+ 0.15
923
+ 0.00
924
+ 0.05
925
+ 0.13
926
+ 1.00 0.52 0.41 0.29 0.31 0.51
927
+ 0
928
+ 0.00
929
+ 0.06
930
+ 0.15
931
+ 0
932
+ 0.08 0.73 0.82 0.82 0.83 0.87 0.90 0.52
933
+ 0.06
934
+ 0.15
935
+ 0.20
936
+ 0.21
937
+ 0.40 0.49
938
+ 1.00
939
+ 0.53 0.36 0.360.70
940
+ 0.00
941
+ 0.06
942
+ 0.06
943
+ 0.00
944
+ 0.06
945
+ 0.20
946
+ 0.96
947
+ 0.07
948
+ 0.01
949
+ 0.02
950
+ 0.01
951
+ 0.07
952
+ 0.14
953
+ 0.25
954
+ 0.29
955
+ 0.07
956
+ 0.08
957
+ 0.15 0.17 0.17 0.33 0.51 0.60 0.61 0.44 0.42 0.41 0.53
958
+ D.61
959
+ 2
960
+ 2
961
+ 1.00 0.95
962
+ 0.00
963
+ 0.00
964
+ 0.00
965
+ 0.00
966
+ 0.05
967
+ 0.05
968
+ 0.35
969
+ 0.37
970
+ 0.06
971
+ 0.12
972
+ 0.05 0.07 0.01 0.21 0.34 0.44 0.45 0.29 0.26 0.29 0.36
973
+ 50.951.00
974
+ 1.00
975
+ 0.62
976
+ 3
977
+ 3
978
+ 0.03
979
+ 0.00
980
+ 0.00
981
+ 0.00
982
+ 0.08
983
+ 0.04
984
+ 0.34
985
+ 0.36
986
+ 0.05
987
+ 0.10
988
+ 0.07 0.08 0.06 0.23 0.36 0.46 0.47 0.31 0.27 0.31 0.36
989
+ 0.96 1.00 1.00 0.60
990
+ 4
991
+ 4
992
+ 0.70 0.61 0.62 0.60
993
+ 0.00
994
+ 0.00
995
+ 0.00
996
+ 0.00
997
+ 0.00
998
+ 0.06
999
+ 0.28
1000
+ 0.30
1001
+ 0.36
1002
+ 0.19 0.18 0.03(
1003
+ 0.27 0.32 0.52 0.36 0.43 0.66
1004
+ 0.51
1005
+ 1.00
1006
+ 5
1007
+ 0.00
1008
+ 5
1009
+ 13 14 15
1010
+ 1
1011
+ 2
1012
+ 3
1013
+ 4
1014
+ 5
1015
+ 6
1016
+ 7
1017
+ 8
1018
+ 9
1019
+ 10
1020
+ 11
1021
+ 12
1022
+ (to)
1023
+ (to)
1024
+ CNV
1025
+ (to)
1026
+ (to)
1027
+ Healthy
1028
+ μm)
1029
+ Drusen
1030
+ NoBiomarker
1031
+ CRORA
1032
+ (250
1033
+ (1000 |
1034
+ (1000μm)
1035
+ RAClustering disease trajectories for biomarker discovery in AMD
1036
+ 11
1037
+ biomarkers. We find two clusters (1-2) describing drusen, five (4-8) describing the
1038
+ transition from drusen to cRORA (250 µm) and three (12-14) describing cRORA
1039
+ (1000 µm). Most healthy samples were in cluster 3, and CNV was spread amongst
1040
+ most late-stage clusters. In Dataset B we find three clusters of healthy looking
1041
+ scans, three of early stages and six in the late stage that capture atrophy.
1042
+ Interpreting candidate biomarkers We now report the results of the single-
1043
+ choice clinical clustering task (described in section 3.4), comparing cluster pairs
1044
+ which are highlighted in blue and pink boxes in Figure 5. In Dataset A, clinicians
1045
+ were easily able to separate clusters 1 and 15 by using known biomarkers, such as
1046
+ hypertransmission and photoreceptor degeneration. The result of this validation
1047
+ experiment was expected, as these clusters were already highly separable under
1048
+ the known set of biomarkers. More interestingly, clinicians were able to exactly
1049
+ differentiate between early AMD clusters 1 and 2 in Dataset A despite their high
1050
+ similarity in known biomarkers. When prompted, all clinicians cite differences
1051
+ in drusen, with two finding differences in the number of small drusen. Their
1052
+ ability to distinguish some of the pairs was mixed, as they sometimes found no
1053
+ consistent or visible differences or had a low inter-rater agreement.
1054
+ 5
1055
+ Discussion and Conclusion
1056
+ In this paper, we proposed a method to automatically discover time-dependent
1057
+ biomarkers that detect periods of disease progression common among groups of
1058
+ patients. By partitioning entire time series into representative sub-trajectories,
1059
+ and then clustering them, we categorised 3,218 total years of disease progression
1060
+ across two datasets. We showed that these clusters rediscovered the established
1061
+ set of OCT biomarkers for AMD, which reinforced the use of our clusters as can-
1062
+ didate biomarkers. Then, by working directly with ophthalmologists, we closed
1063
+ the loop in our automated biomarker discovery. To this end ophthalmologists
1064
+ compared clusters that were indistinguishable using current grading systems,
1065
+ yet were separable in contrastive feature space. We envision that further inves-
1066
+ tigation into sub-trajectory clusters could advance understanding of how AMD
1067
+ progresses, and potentially lead to grading systems with greater prognostic value.
1068
+ Our method is applicable to any dataset studying any disease with time series
1069
+ of images. While our method identified two clusters that described drusen in
1070
+ Dataset A and three that described healthy-looking scans in Dataset B, most
1071
+ clusters were associated with intermediate and late-stage AMD. This is due
1072
+ to the overrepresentation of patients with late disease in our datasets. In or-
1073
+ der to find more clusters categorising progression in early AMD, we aim to
1074
+ repeat this analysis in datasets that begin imaging patients earlier in their over-
1075
+ all progression. Moreover, due to the slow progression of AMD, a large number
1076
+ of sub-trajectories captured unchanging disease states. In order to isolate sub-
1077
+ trajectories capturing the periods of greatest disease progression, we intend to
1078
+ increase the number of clusters K and interpret those that convert to late disease
1079
+ the fastest.
1080
+
1081
+ 12
1082
+ Holland et al.
1083
+ Conclusion Inspired by inadequate grading systems for disease progression
1084
+ in early AMD, we proposed the first method to analyse disease progression as
1085
+ clustered trajectories in self-supervised feature space. By correlating our clusters
1086
+ with known OCT biomarkers for AMD, we reinforced their potential as time-
1087
+ dependent biomarkers for disease progression. After this, we closed the loop on
1088
+ automated biomarker discovery by working directly with ophthalmologists to
1089
+ investigate our clusters. We envision that self-supervised learning can enable
1090
+ detection of patterns of disease progression in time series of patient populations,
1091
+ which can lead to grading systems with greater prognostic value.
1092
+ References
1093
+ 1. Bian, J., et al.: A survey on trajectory clustering analysis. CoRR abs/1802.06971
1094
+ (2018), http://arxiv.org/abs/1802.06971
1095
+ 2. Bird, A.C., et al.: An international classification and grading system for age-
1096
+ related maculopathy and age-related macular degeneration. Survey of ophthal-
1097
+ mology 39(5), 367–374 (1995)
1098
+ 3. Chen, K.G., et al.: Longitudinal study of dark adaptation as a functional outcome
1099
+ measure for age-related macular degeneration. Ophthalmology 126(6), 856–865
1100
+ (2019)
1101
+ 4. Chen, T., et al.: A simple framework for contrastive learning of visual represen-
1102
+ tations. In: International conference on machine learning. pp. 1597–1607. PMLR
1103
+ (2020)
1104
+ 5. Dmitrenko, A., et al.: Self-supervised learning for analysis of temporal and mor-
1105
+ phological drug effects in cancer cell imaging data. In: Medical Imaging with Deep
1106
+ Learning (2021)
1107
+ 6. Ferreira, N., et al.: Vector field k-means: Clustering trajectories by fitting multiple
1108
+ vector fields. In: Computer Graphics Forum. vol. 32, pp. 201–210. Wiley Online
1109
+ Library (2013)
1110
+ 7. Ferris, F.L., et al.: A simplified severity scale for age-related macular degeneration.
1111
+ Archives of ophthalmology 123(11), 1570–1574 (2005)
1112
+ 8. Ferris III, F.L., Wilkinson, C., Bird, A., Chakravarthy, U., Chew, E., Csaky, K.,
1113
+ Sadda, S.R., for Macular Research Classification Committee, B.I., et al.: Clinical
1114
+ classification of age-related macular degeneration. Ophthalmology 120(4), 844–851
1115
+ (2013)
1116
+ 9. Grill, J.B., et al.: Bootstrap your own latent-a new approach to self-supervised
1117
+ learning. NeurIPS 33, 21271–21284 (2020)
1118
+ 10. Holland, R., et al.: Metadata-enhanced contrastive learning from retinal optical
1119
+ coherence tomography images. CoRR abs/2208.02529 (2022)
1120
+ 11. Joachim, N., et al.: Incidence and progression of reticular drusen in age-related
1121
+ macular degeneration: findings from an older australian cohort. Ophthalmology
1122
+ 121(4), 917–925 (2014)
1123
+ 12. Klein, R., et al.: Harmonizing the classification of age-related macular degeneration
1124
+ in the three-continent amd consortium 21(1), 14–23 (2014)
1125
+ 13. Lee, J.G., et al.: Trajectory clustering: a partition-and-group framework. In: ACM
1126
+ SIGMOD. pp. 593–604 (2007)
1127
+ 14. Mitchell, P., et al.: Age-related macular degeneration. The Lancet 392(10153),
1128
+ 1147–1159 (2018)
1129
+
1130
+ Clustering disease trajectories for biomarker discovery in AMD
1131
+ 13
1132
+ 15. Sadda, S.R., et al.: Consensus definition for atrophy associated with age-related
1133
+ macular degeneration on oct: classification of atrophy report 3. Ophthalmology
1134
+ 125(4), 537–548 (2018)
1135
+ 16. Schlanitz, F.G., et al.: Drusen volume development over time and its relevance to
1136
+ the course of age-related macular degeneration. BJO 101(2), 198–203 (2017)
1137
+ 17. Schlegl, T., et al.: f-anogan: Fast unsupervised anomaly detection with generative
1138
+ adversarial networks. Medical image analysis 54, 30–44 (2019)
1139
+ 18. Seeböck, P., et al.: Unsupervised identification of disease marker candidates in
1140
+ retinal oct imaging data. IEEE TMI 38(4), 1037–1047 (2018)
1141
+ 19. Steinberg, J.S., et al.: Longitudinal analysis of reticular drusen associated with
1142
+ geographic atrophy in age-related macular degeneration. IOVS 54(6), 4054–4060
1143
+ (2013)
1144
+ 20. Vogel, J.W., et al.: Four distinct trajectories of tau deposition identified in
1145
+ alzheimer’s disease. Nature medicine 27(5), 871–881 (2021)
1146
+ 21. Waldstein, S.M., et al.: Unbiased identification of novel subclinical imaging
1147
+ biomarkers using unsupervised deep learning. Scientific reports 10(1), 1–9 (2020)
1148
+ 22. Wong, W.L., et al.: Global prevalence of age-related macular degeneration and dis-
1149
+ ease burden projection for 2020 and 2040: a systematic review and meta-analysis.
1150
+ The Lancet Global Health 2(2), e106–e116 (2014)
1151
+ 23. Young, A.L., et al.: A data-driven model of biomarker changes in sporadic
1152
+ alzheimer’s disease. Brain 137(9), 2564–2577 (2014)
1153
+ 24. Zhao, A., et al.: Prognostic imaging biomarker discovery in survival analysis for id-
1154
+ iopathic pulmonary fibrosis. In: MICCAI proceedings. pp. 223–233. Springer (2022)
1155
+ 25. Zheng, Q., et al.: Pathological cluster identification by unsupervised analysis in
1156
+ 3,822 uk biobank cardiac mris. Front. cardiovasc. med 7, 539788 (2020)
1157
+
Y9E3T4oBgHgl3EQfcgoK/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
Z9E5T4oBgHgl3EQfDg58/content/tmp_files/2301.05406v1.pdf.txt ADDED
@@ -0,0 +1,1689 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Non-Gaussianity in the cosmic microwave background from loop quantum
2
+ cosmology
3
+ Roshna K∗ and V. Sreenath†
4
+ Department of Physics, National Institute of Technology Karnataka, Surathkal, Mangaluru 575025, India.
5
+ Primordial non-Gaussianity has set strong constraints on models of the early universe.
6
+ Studies have shown that Loop Quantum Cosmology (LQC), which is an attempt to extend
7
+ inflationary scenario to planck scales, leads to a strongly scale dependent and oscillatory non-
8
+ Gaussianity. In particular, the non-Gaussianity function fNL(k1, k2, k3) generated in LQC,
9
+ though similar to that generated during slow roll inflation at small scales, is highly scale
10
+ dependent and oscillatory at large wavelengths. In this work, we investigate the imprints of
11
+ such a primordial bispectrum in the bispectrum of Cosmic Microwave Background (CMB).
12
+ Inspired by earlier works, we propose an analytical template for the primordial bispectrum
13
+ in LQC and compute the corresponding reduced bispectra of temperature and electric po-
14
+ larisation and their three-point cross-correlations. We show that CMB bispectra generated
15
+ in LQC is consistent with the observations from Planck. We conclude with a discussion of
16
+ our results and its implications to LQC.
17
+ I.
18
+ INTRODUCTION
19
+ Numerous theoretical insights together with several observational efforts, spanned over a cen-
20
+ tury, have enabled us to arrive at a compelling model of our Universe referred to as the standard
21
+ model or the Lambda Cold Dark Matter (ΛCDM) model [1]. According to this model, the seeds
22
+ of the current distribution of galaxies spread over the fabric of spacetime known as the large scale
23
+ structure were sown during the earliest phase of the universe. Tiny perturbations generated in the
24
+ early universe lead to tiny anisotropies in the Cosmic Microwave Background (CMB) which in turn
25
+ lead to the inhomegeneous large scale distribution of galaxies that we see today. Though we have
26
+ a good level of understanding of this evolution, several details are yet to be worked out. One such
27
+ detail concerns the origin of these perturbations in our Universe.
28
+ Inflation, see, for instance, [2–5], due to its simplicity, provides the most popular explanation
29
+ for the origin of these perturbations [6, 7] (For a discussion on alternate views, see [8, 9].). In
30
+ inflationary scenario, quantum fluctuations in the inflaton leads to the primordial perturbations.
31
+ Appealing to the nearly de Sitter symmetry of the spacetime during inflation, we assume that at
32
+ a time when the perturbations are sufficiently sub-horizon, quantum perturbations are generated
33
+ in the Bunch-Davies vacuum. Such a prescription has been highly successful, in that, it leads to
34
+ primordial perturbations that are nearly Gaussian and scale invariant as demanded by observations
35
+ [1, 6, 10]. Even though inflation is successful, it is still an incomplete theory. We do not take in to
36
+ account the evolution of perturbations before the time at which the initial conditions are imposed.
37
+ In fact, inflation does not account for the physics in the planck regime close to the big bang
38
+ singularity. There have been several attempts to address these issues. In this work, we will concern
39
+ ourselves with loop quantum cosmology (LQC) [11–15].
40
+ Loop quantum cosmology is an attempt to extend inflationary scenario to the planck regime
41
+ using principles of loop quantum gravity [13–19]. In LQC, quantum gravitational effects in the
42
+ planck regime leads to a quantum bounce [11, 12]. Thus in LQC, a quantum bounce precedes
43
+ the inflationary phase. Generation and evolution of perturbations in LQC have been extensively
44
45
46
+ arXiv:2301.05406v1 [astro-ph.CO] 13 Jan 2023
47
+
48
+ 2
49
+ studied at the level of primordial power spectra [20–46] and primordial non-Gaussianity [47–50]. In
50
+ general, studies show that the effect of the bounce is to introduce an additional scale corresponding
51
+ to the curvature at the bounce. Modes of perturbations which have comparable length to this new
52
+ scale gets modified leading to a highly scale dependent power spectrum. At smaller wavelengths, the
53
+ perturbations are not affected by the bounce and the power spectrum is nearly scale invariant as in
54
+ slow roll inflation [29]. Perturbations show a similar behaviour at second order in perturbations [47,
55
+ 49]. Studies show that primordial non-Gaussianity quantified using the function fNL(k1, k2, k3),
56
+ at scales comparable to the curvature at the bounce, is strongly scale dependent and oscillatory
57
+ with a very large amplitude. At smaller scales, the fNL(k1, k2, k3) is similar to that in slow roll.
58
+ Studies also show that the bispectrum is more sensitive to the bounce than the power spectrum.
59
+ Assuming sixty or so e-folds of inflation, the scale at which the imprints of the bounce, on
60
+ primordial perturbations, occur depends on the amount of expansion between the bounce and the
61
+ onset of inflation. Observational constraints from the CMB temperature power spectrum demand
62
+ that any departure from scale invariance should happen only at multipoles of ℓ ≲ 30 [1, 51]. If we
63
+ assume that, the effects of primordial power spectrum on the CMB is observable at ℓ ≲ 30, then,
64
+ since the bispectrum is more sensitive to the effects of the bounce than the power spectrum [47, 49],
65
+ there is a possibility that the imprints of large, scale dependent and oscillatory primordial non-
66
+ Gaussianity is observable at larger multipoles. Hence it is important to investigate the consistency
67
+ of LQC with observations by Planck. With this motivation, in this work, we compute the imprints
68
+ of such a non-Gaussianity in the temperature (T) and electric polarisation (E) of the CMB. We
69
+ assume an analytical template for primordial non-Gaussianity generated in LQC, compute the
70
+ ⟨TTT⟩, ⟨TTE⟩, ⟨TEE⟩ and ⟨EEE⟩ correlations and show that they are similar to those generated
71
+ in slow roll inflation and hence is consistent with observations by Planck.
72
+ The rest of the paper is organised as follows. In the next section, we briefly introduce the essen-
73
+ tials of LQC and present analytical templates for the primordial power spectrum and bispectrum.
74
+ In section III, we discuss the essential formulae to compute the three-point correlation functions
75
+ of anisotropies in temperature and electric polarisation. In section IV, we apply these formulae to
76
+ LQC. We present the numerical techniques and our calculation of reduced bispectra of tempera-
77
+ ture fluctuations and electric polarisation and their three-point cross-correlations in section V. We
78
+ conclude the paper with a summary and discussion of our results and their consequences to LQC
79
+ in section VI.
80
+ II.
81
+ LOOP QUANTUM COSMOLOGY
82
+ In this section, we will discuss the essentials of LQC that is relevant to this paper (for reviews,
83
+ see, for instance, [13–15]). In particular, we will discuss LQC as applied to FLRW geometries
84
+ sourced by a scalar field φ and scalar perturbations δφ(⃗x) living on this background.
85
+ A.
86
+ Background
87
+ In LQC, FLRW background geometry is described by a wavefunction ΨFLRW(v, φ), which satis-
88
+ fies the equation ˆHFLRWΨFLRW(v, φ) = 0, where ˆHFLRW is the Hamiltonian operator corresponding
89
+ to the classical background Hamiltonian and v is the volume factor which is proportional to the
90
+ cube of scale factor a. Numerical investigations of such a system has shown that the scale factor
91
+ undergoes a bounce [11, 12, 52, 53]. It turns out, if the wave function is sharply peaked over the
92
+ values of scale factor, the behaviour of scale factor can be described by certain effective equations
93
+
94
+ 3
95
+ −104
96
+ −102
97
+ 0
98
+ 102
99
+ 104
100
+ 106
101
+ t (TPl)
102
+ 103
103
+ 108
104
+ 1013
105
+ 1018
106
+ 1023
107
+ 1028
108
+ 1033
109
+ a(t)
110
+ Inflation
111
+ −10
112
+ −5
113
+ 0
114
+ 5
115
+ 10
116
+ 0
117
+ 2
118
+ 4
119
+ 6
120
+ 8
121
+ 100
122
+ 101
123
+ 102
124
+ 103
125
+ 104
126
+ 105
127
+ 106
128
+ 107
129
+ t (TPl)
130
+ −5
131
+ 0
132
+ 5
133
+ 10
134
+ 15
135
+ φ(t)
136
+ FIG. 1.
137
+ Figure illustrates the behaviour of scale factor (left) and scalar field(right) in LQC. As mentioned
138
+ in the text, scale factor undergoes a bounce preceding inflation. Scalar field starts rolling up the potential
139
+ until its kinetic energy becomes zero and then starts slowly rolling down the potential leading to inflation.
140
+ In making this plot, we have worked with the mass of scalar field to be consistent with the constraints on
141
+ the amplitude of the primordial power spectrum and with ρsup = 0.41m4
142
+ Pl.
143
+ [11, 12, 30, 52, 53], namely
144
+ � ˙a
145
+ a
146
+ �2
147
+ = κ
148
+
149
+
150
+ 1 −
151
+ ρ
152
+ ρsup
153
+
154
+ ,
155
+ ¨a
156
+ a = −κ
157
+ 6 ρ
158
+
159
+ 1 − 4
160
+ ρ
161
+ ρsup
162
+
163
+ − κ
164
+ 2 P
165
+
166
+ 1 − 2
167
+ ρ
168
+ ρsup
169
+
170
+ ,
171
+ (2.1)
172
+ where ρ, P are the energy density and pressure of the scalar field and κ = 8 π G. From the above
173
+ expression, it is clear that at ρ = ρsup, Hubble parameter H = ˙a/a = 0 and ¨a/a > 0 i.e.
174
+ scale
175
+ factor is at minimum. In other words, the universe undergoes a bounce at ρ = ρsup. Further, if
176
+ we assume that the scalar field is governed by a potential V (φ), then the evolution of scalar field
177
+ is given by
178
+ ¨φ + 3 H ˙φ + Vφ = 0,
179
+ (2.2)
180
+ where Vφ = dV/dφ. For a suitable potential, inflationary phase will set in after the bounce [34, 54–
181
+ 58]. The background dynamics in LQC with a scalar field governed by a quadratic potential is
182
+ illustrated in Figure 1.
183
+ B.
184
+ Perturbations
185
+ We will follow dressed metric approach to describe primordial perturbations in LQC [22–24, 29,
186
+ 47, 49]. In this approach, we assume that the wavefunction takes the form Ψ = ΨFLRW(v, φ) ⊗
187
+ δΨ(v, φ, δφ), which satisfies the equation ˆHΨ = 0, where ˆH = ˆHFLRW + ˆHpert. As mentioned
188
+ earlier, ΨFLRW(v, φ) satisfies the equation ˆHFLRWΨFLRW(v, φ) = 0. Perturbations are treated as
189
+ test fields living on the background FLRW geometries described by ΨFLRW(v, φ). In practice, this
190
+ implies that perturbations can be evolved using the classical Hamiltonian but with the background
191
+ functions in them described by the effective equations. This is similar to perturbations living as test
192
+ fields on a curved space time described by a ‘dressed’ metric which satisfies the effective equations.
193
+
194
+ 4
195
+ In order to compute primordial bispectrum, we need to consider Hamiltonian up to third order
196
+ in perturbations, i.e.
197
+ we need Hpert = H(2) + H(3). There are two approaches to arrive at the
198
+ Hamiltonian describing perturbations, one can either use gauge invariant variables or rather work
199
+ with a fixed gauge. We follow the latter approach. In particular, we will work with spatially flat
200
+ gauge [47, 59].
201
+ The second order Hamiltonian describing perturbations δφ in the spatially flat gauge is
202
+ H(2) =
203
+
204
+ d3x N S(2)(⃗x) = N 1
205
+ 2
206
+
207
+ d3x
208
+ � 1
209
+ a3 δpφ2 + a3 (∂δφ)2 + a3 U δφ2
210
+
211
+ ,
212
+ (2.3)
213
+ with the potential U given by
214
+ U = −9
215
+ p4
216
+ φ
217
+ a8π2a
218
+ + 3
219
+
220
+ p2
221
+ φ
222
+ a6 − 6 pφ
223
+ a πa
224
+ Vφ + Vφφ + 6 pφ ˙pφ
225
+ a4 πa
226
+ − 3
227
+ p2
228
+ φ ˙πa
229
+ a4 π2a
230
+ − 3
231
+ ˙a p2
232
+ φ
233
+ a5 πa
234
+ .
235
+ (2.4)
236
+ In the above expressions, πa, pφ and δpφ are momenta conjugate to a, φ and δφ respectively. Setting
237
+ lapse N = 1 will imply cosmic time and N = a corresponds to conformal time. Hamiltonian at
238
+ third order in perturbations is
239
+ H(3) = N
240
+
241
+ d3x
242
+ ��
243
+ 9 κ p3
244
+ φ
245
+ 4 a4 πa
246
+
247
+ 27 p5
248
+ φ
249
+ 2 a6π3a
250
+ − 3 a2 pφ Vφφ
251
+ 2 πa
252
+ + a3 Vφφφ
253
+ 6
254
+
255
+ δφ3
256
+
257
+ 3 pφ
258
+ 2 a4 πa
259
+ δp2
260
+ φ δφ −
261
+ 9 p3
262
+ φ
263
+ a5π2a
264
+ δpφδφ2 − 3 a2 pφ
265
+ 2 πa
266
+ δφ (⃗∂δφ)2 +
267
+ 3 p2
268
+ φ
269
+ N a πa
270
+ δφ2∂2χ + 3
271
+ 2
272
+ a2 pφ
273
+ N2 κ πa
274
+ δφ ∂2χ ∂2χ
275
+ + 3
276
+ p2
277
+ φ
278
+ N a πa
279
+ δφ ∂iχ∂iδφ + 1
280
+ N δpφ ∂iδφ ∂iχ − 3
281
+ 2
282
+ a2 pφ
283
+ N2 κ πa
284
+ δφ ∂i∂jχ ∂i∂jχ
285
+
286
+ ,
287
+ (2.5)
288
+ where ∂2χ = (−3 N κ/a)
289
+ ��
290
+
291
+ 2 − a5 Vφ
292
+ κ πa
293
+
294
+ δφ −
295
+
296
+ κ a πa δpφ
297
+
298
+ .
299
+ From the second order Hamiltonian H(2), one can derive the free evolution of the scalar pertur-
300
+ bation, given by,
301
+ (□ − U(t)) δφ(⃗x, t) = 0,
302
+ (2.6)
303
+ where □ is the d’Alembertian of the FLRW background metric. The third order Hamiltonian H(3)
304
+ provides the self-interaction of the scalar perturbations.
305
+ The perturbations, since they evolve through the bounce and then through the inflationary
306
+ phase, carry signatures of the early universe which they imprint on the CMB. Perturbations are
307
+ quantified using correlation functions.
308
+ In order to compute correlation functions, one need to
309
+ promote δφ to an operator ˆδφ. The field operator ˆδφ is then expanded in terms of annihilation
310
+ and creation operators as
311
+ ˆδφ(⃗x, η) =
312
+
313
+ d3k
314
+ (2π)3 ˆδφ⃗k(η) ei⃗k·⃗x =
315
+
316
+ d3k
317
+ (2π)3
318
+
319
+ ˆA⃗k ϕk(η) + ˆA†
320
+ −⃗k ϕ∗
321
+ k(η)
322
+
323
+ ei⃗k·⃗x
324
+ (2.7)
325
+ where [ ˆA⃗k, ˆA†
326
+ ⃗k′] = ℏ (2π)3 δ(3)(⃗k + ⃗k′), [ ˆA⃗k, ˆA⃗k′] = 0 = [ ˆA†
327
+ ⃗k, ˆA†
328
+ ⃗k′]. The mode functions ϕk(η) satisfy
329
+ the equation
330
+ ϕ′′
331
+ k + 2a′
332
+ a ϕ′
333
+ k + (k2 + a2 U) ϕk = 0 ,
334
+ (2.8)
335
+
336
+ 5
337
+ where k2 ≡ kikj δij is the comoving wavenumber, and prime indicates derivative with respect to
338
+ conformal time. The scalar power spectrum of ˆδφ is a dimensionless function that quantifies the
339
+ two-point correlation in momentum space via
340
+ ⟨0| ˆδφ⃗k(η) ˆδφ⃗k′(η)|0⟩ ≡ (2π)3δ(3)(⃗k + ⃗k′)2π2
341
+ k3 Pδφ(k, η) ,
342
+ (2.9)
343
+ where |0⟩ is the vacuum annihilated by the operators ˆA⃗k for all ⃗k. Power spectrum, in terms of
344
+ mode functions, is Pδφ(k, η) = (ℏ k3/2π2) |ϕk(η)|2.
345
+ The three-point function of ˆδφ at tree level is given by [47, 59]
346
+ ⟨ ˆδφ⃗k1(η) ˆδφ⃗k2(η) ˆδφ⃗k3(η) ⟩ = − i/ℏ
347
+
348
+ dη′⟨0|
349
+
350
+ ˆδφ
351
+ I
352
+ ⃗k1(η) ˆδφ
353
+ I
354
+ ⃗k2(η) ˆδφ
355
+ I
356
+ ⃗k3(η), ˆHI
357
+ int(η′)
358
+
359
+ |0⟩ + O(H2
360
+ int),
361
+ (2.10)
362
+ where ˆHI
363
+ int(η) is the operator corresponding to H(3) in the interaction picture.
364
+ Even though we worked in spatially flat gauge, it is convenient to compute correlation functions
365
+ in terms of curvature perturbations R. This is because, curvature perturbations have a unique
366
+ property that they stop evolving after they cross the horizon and remain constant till they re-enter
367
+ horizon towards late radiation domination or during early matter domination epoch, saving us a
368
+ lot of computational time. Curvature perturbations are related to perturbations in scalar field
369
+ through the relation [47, 59]
370
+ R(⃗x, η) = −a
371
+ z δφ(⃗x, η) +
372
+
373
+ −3
374
+ 2 + 3 Vφ a5
375
+ κ Pφ πa
376
+ + κ
377
+ 4
378
+ z2
379
+ a2
380
+ � �a
381
+ z δφ(⃗x, η)
382
+ �2
383
+ + · · · ,
384
+ (2.11)
385
+ where trailing dots indicates terms that leads to subdominant terms in the three-point functions
386
+ when evaluated towards the end of inflation.
387
+ The power spectrum of curvature perturbation is related to that of scalar modes ˆδφ⃗k(η) through
388
+ the relation
389
+ PR(k) =
390
+ �a(ηend)
391
+ z(ηend)
392
+ �2
393
+ Pδφ(k, η),
394
+ (2.12)
395
+ where z = −6 pφ/(κ πa).
396
+ The three-point function of curvature perturbation can be obtained in terms of ˆδφ⃗k(η) by using
397
+ Eq. (2.11) as
398
+ ⟨0| ˆR⃗k1 ˆR⃗k2 ˆR⃗k3|0⟩ =
399
+
400
+ −a
401
+ z
402
+ �3
403
+ ⟨0| ˆδφ⃗k1 ˆδφ⃗k2 ˆδφ⃗k3|0⟩
404
+ +
405
+
406
+ −3
407
+ 2 + 3 Vφ a5
408
+ κ pφ πa
409
+ + κ
410
+ 4
411
+ z2
412
+ a2
413
+ � �
414
+ −a
415
+ z
416
+ �4 � �
417
+ d3p
418
+ (2π)3 ⟨0| ˆδφ⃗k1 ˆδφ⃗k2 ˆδφ⃗p ˆδφ⃗k3−⃗p|0⟩ + (⃗k1 ↔ ⃗k3) + (⃗k2 ↔ ⃗k3)
419
+ + · · ·
420
+
421
+ .
422
+ (2.13)
423
+ The wave numbers of three modes in the three-point function are constrained by a Dirac delta
424
+ function. We define the scalar bispectrum as the three-point function sans Dirac delta function as
425
+ ⟨0| ˆR⃗k1 ˆR⃗k2 ˆR⃗k3|0⟩ ≡ (2π)3δ(3)(⃗k1 + ⃗k2 + ⃗k3) BR(k1, k2, k3) .
426
+ (2.14)
427
+ The amplitude of bispectrum can be quantified using a dimensionless function fNL(k1, k2 , k3), akin
428
+ to the dimensionless power spectrum PR(k) that quantifies two-point correlations, as
429
+ fNL(k1, k2, k3) ≡ −5
430
+ 6BR(k1, k2, k3) × (∆k1∆k2 + ∆k1∆k3 + ∆k2∆k3)−1
431
+ (2.15)
432
+ where ∆k ≡ 2 π2
433
+ k3 PR(k).
434
+
435
+ 6
436
+ −104 −103 −102 −101
437
+ −1000 100
438
+ 101
439
+ 102
440
+ 103
441
+ 104
442
+ 105
443
+ 106
444
+ t (TPl)
445
+ 10−5
446
+ 10−3
447
+ 10−1
448
+ 101
449
+ 103
450
+ 105
451
+
452
+ |Ω(η)| (MPl)
453
+ k⋆
454
+ kLQC
455
+ kI
456
+ FIG. 2.
457
+ The figure represents the relevant scales in LQC. kLQC is the scale corresponding to the value of
458
+ curvature at the bounce. kI corresponds to the smallest scale that is sub-horizon during inflation. As can
459
+ be seen, only modes kLQC ≳ k > kI are excited during the bounce and hence are scale dependent. Modes
460
+ with larger wavenumbers are excited only during horizon crossing towards the end of inflation and hence
461
+ will be scale invariant.
462
+ C.
463
+ Templates of scalar power spectrum and bispectrum
464
+ In order to understand the evolution of perturbations in LQC, let us rewrite Eqn. (2.8),
465
+ v′′
466
+ k +
467
+
468
+ k2 + Ω(η)
469
+
470
+ vk = 0,
471
+ (2.16)
472
+ where vk = a ϕk is the Mukhanov-Sasaki variable and Ω(η) = a2 U −
473
+ a′′
474
+ a .
475
+ We compare the
476
+ behaviour of
477
+
478
+ |Ω(η)| as a function of time with relevant wavenumbers in figure 2. As shown in the
479
+ figure, all observationally relevant wavenumbers are adiabatic much before the bounce and hence
480
+ we can impose adiabatic initial conditions. From the figure, it is also clear that there are two
481
+ relevant scales in the problem. The value of curvature at the bounce defines a scale kLQC and the
482
+ value of curvature at the onset of inflation defines a scale kI. Wavenumbers which are much larger
483
+ than kLQC, are not effected by the bounce and they will be in Bunch-Davies vacuum at the onset
484
+ of inflation. This implies that power spectrum of modes k >> kLQC will be nearly scale invariant
485
+ as in slow roll inflation. Modes which are comparable to kLQC and larger than kI will be excited
486
+ both during the bounce as well as during the horizon exit during inflation. These modes are in
487
+ excited non-Gaussian states during the onset of inflation and hence they will be further amplified
488
+ as they exit the horizon during inflation. Hence, the modes kI < k < kLQC will be strongly scale
489
+ dependent. Modes whose wavenumbers are smaller than kI are always superhorizon and hence
490
+ they are never excited. The primordial power spectrum and bispectrum are evaluated towards the
491
+ end of inflation when all the relevant modes are well outside the horizon.
492
+ The primordial power spectrum and bispectrum can be calculated numerically.
493
+ Given the
494
+ background dynamics described in Eqns. (2.1, 2.2), the evolution of perturbations are found by
495
+
496
+ 7
497
+ 10−7
498
+ 10−6
499
+ 10−5
500
+ 10−4
501
+ 10−3
502
+ 10−2
503
+ k
504
+
505
+ Mpc−1�
506
+ 10−10
507
+ 10−9
508
+ 10−8
509
+ 10−7
510
+ 10−6
511
+ PR(k)
512
+ kLQC
513
+ k⋆
514
+ analytical result
515
+ numerical result
516
+ 10−5
517
+ 10−4
518
+ 10−3
519
+ 10−2
520
+ k
521
+
522
+ Mpc−1�
523
+ −106
524
+ −105
525
+ −104
526
+ −103
527
+ −102
528
+ −101
529
+ −1000
530
+ 100
531
+ 101
532
+ 102
533
+ 103
534
+ 104
535
+ 105
536
+ 106
537
+ fNL(k, k, k)
538
+ kLQC
539
+ k⋆
540
+ analytical result
541
+ numerical result
542
+ FIG. 3.
543
+ The primordial power spectrum and the non-Gaussianity function generated in LQC obtained
544
+ numerically (in black). Analytical templates for power spectrum and non-Gaussianity given in Eqns. (2.17)
545
+ and (2.19) (in grey).
546
+ solving Eqn. (2.8). The power spectrum of curvature perturbation can then be calculated using
547
+ Eqn. (2.12). Calculation of ⟨ ˆδφ⃗k1(η) ˆδφ⃗k2(η) ˆδφ⃗k3(η) ⟩ requires one to perform integrals in Eqn.
548
+ (2.10). The ⟨0| ˆR⃗k1 ˆR⃗k2 ˆR⃗k3|0⟩ three-point function of curvature perturbation can then be calculated
549
+ using Eqn. (2.13). The dimensionless non-Gaussianity function of curvature perturbation is then
550
+ calculated by using Eqn. (2.15). This numerical calculation of primordial power spectrum and
551
+ non-Gaussianity has been implemented in class_lqc [47, 49]. We present the results obtained
552
+ using that code in figure 3.
553
+ For calculating the three-point functions involving temperature and electric polarisation, one
554
+ needs to convolve the primordial bispectrum with the CMB transfer functions. For performing
555
+ this calculation, it is convenient to have analytical templates of primordial power spectrum and
556
+ bispectrum. Following [46, 60, 61], we will use the following template for describing the power
557
+ spectrum. It is given by
558
+ PR(k) = As
559
+
560
+
561
+
562
+
563
+
564
+
565
+
566
+
567
+
568
+ ( k
569
+ kI )2(
570
+ kI
571
+ kLQC )q
572
+ if k ≤ kI,
573
+ (
574
+ k
575
+ kLQC )q
576
+ if kI < k ≤ kLQC,
577
+ (
578
+ k
579
+ kLQC )(ns−1) if k > kLQC,
580
+ (2.17)
581
+ where we work with q = −0.7, kI = 5 × 10−5 k⋆, kLQC = 0.1 k⋆ and k⋆ = 0.002Mpc−1 represents
582
+ the pivot scale. The amplitude of power spectrum As and the spectral index ns have been set to
583
+ their values obtained by Planck. The analytical template for power spectra is drawn along with
584
+ the exact numerical calculation in figure 3.
585
+ As is evident from figure 3, the primordial non-Gaussianity fNL(k1, k2, k3) is scale dependent
586
+ and oscillatory. The exponential decay in the value of fNL(k1, k2, k3) as k ≈ kLQC was explained
587
+ in [47, 49] by analysing the poles of the integrand in Eqn. (2.10). In particular, by analysing the
588
+ pole of scale factor around the bounce, the analytical behaviour of fNL sans oscillations was found
589
+ to be
590
+ fNL(k1, k2, k3) ∝ e
591
+ −α k1 + k2 + k3
592
+ kb
593
+ ,
594
+ (2.18)
595
+ where α = 0.647. To the above scale dependent form, we incorporate the oscillations and also add
596
+ the fact that for k > kb the shape of bispectrum approaches that of slow roll. Thus, we obtain the
597
+
598
+ 8
599
+ analytical template for LQC to be
600
+ fNL(k1, k2, k3) = f
601
+ bounce
602
+ NL
603
+ e
604
+ −α k1+k2+k3
605
+ kb
606
+ sin
607
+ �k1 + k2 + k3
608
+ kI
609
+
610
+ + f
611
+ loc
612
+ NL.
613
+ (2.19)
614
+ The above analytical template is plotted along with the exact numerical of fNL(k, k, k) result in
615
+ figure 3, where we have worked with kb = 1.5 kLQC, f
616
+ loc
617
+ NL = 10−2 and f
618
+ bounce
619
+ NL
620
+ = 80000. The value
621
+ of f
622
+ loc
623
+ NL that we work with is similar to that produced in slow roll inflation. As is evident, from the
624
+ figure, the template qualitatively captures the essential features of the primordial non-Gaussianity
625
+ in LQC.
626
+ III.
627
+ CMB BISPECTRA
628
+ Primordial perturbations leave their imprints in the CMB radiation as temperature fluctuations
629
+ and as electric and magnetic polarisations, often referred to as E and B modes respectively (see,
630
+ for instance, [62–64]). The temperature fluctuations and E modes are produced from primordial
631
+ scalar perturbations, whereas B modes are not.
632
+ Since we are interested in understanding the
633
+ imprints of scalar bispectrum, we will focus on the bispectra of temperature fluctuations and
634
+ electric polarisation and their three-point cross-correlations. In this section, we will discuss the
635
+ essential aspects of computing these bispectra.
636
+ Since CMB is observed on a sphere, namely the surface of last scattering, it is convenient to
637
+ decompose it in terms of spherical harmonics,
638
+ X(ˆn) =
639
+
640
+ ℓ,m
641
+ aX
642
+ ℓm Yℓm(ˆn)
643
+ (3.1)
644
+ where X could be either fluctuation in temperature defined as (T(ˆn) − ¯T)/ ¯T, where ¯T is the
645
+ mean temperature of the CMB, or electric polarisation E(ˆn). The multipole aX
646
+ ℓm corresponding to
647
+ anisotropies in the temperature and electric polarisation is related to the curvature perturbation
648
+ through the relation
649
+ aX
650
+ ℓm = 4π (−i)ℓ
651
+
652
+ d3k
653
+ (2π)3 Rk ∆X
654
+ ℓ (k) Yℓm(k).
655
+ (3.2)
656
+ In the above, ∆X
657
+ ℓ is the transfer function which captures the physics post horizon exit of perturba-
658
+ tions towards the end of inflation. We are interested in calculating the three-point function of these
659
+ multipoles of the form ⟨aX
660
+ ℓ1m1 aY
661
+ ℓ2m2 aZ
662
+ ℓ3m3⟩, where X, Y and Z can be either temperature fluctuations
663
+ or E mode polarisation and where the average is over different realisations of the Universe.
664
+ The three-point function of multipole coefficients can be expressed in terms of three-point
665
+ functions of primordial perturbations as [62, 65–68]
666
+ ⟨aX
667
+ ℓ1m1 aY
668
+ ℓ2m2 aZ
669
+ ℓ3m3 ⟩ = (4π)3 (−i)ℓ1+ℓ2+ℓ3
670
+
671
+ d3k1
672
+ (2π)3
673
+
674
+ d3k2
675
+ (2π)3
676
+
677
+ d3k3
678
+ (2π)3 ∆X
679
+ ℓ1∆Y
680
+ ℓ2∆Z
681
+ ℓ3
682
+ × ⟨Rk1Rk2Rk3⟩ Yℓ1m1(ˆk1) Yℓ2m2(ˆk2) Yℓ3m3(ˆk3).
683
+ (3.3)
684
+ Using Eqn. (2.14) and expressing the Dirac-Delta function in its exponential form, we obtain
685
+ ⟨aX
686
+ ℓ1m1 aY
687
+ ℓ2m2 aZ
688
+ ℓ3m3 ⟩ = (4π)3 (−i)ℓ1+ℓ2+ℓ3
689
+
690
+ d3k1
691
+ (2π)3
692
+
693
+ d3k2
694
+ (2π)3
695
+
696
+ d3k3
697
+ (2π)3 ∆X
698
+ ℓ1∆Y
699
+ ℓ2∆Z
700
+ ℓ3
701
+ ×
702
+
703
+ d3x ei(⃗k1 +⃗k2 +⃗k3).⃗x BR(k1, k2, k3) Yℓ1m1(ˆk1) Yℓ2m2(ˆk2) Yℓ3m3(ˆk3). (3.4)
704
+
705
+ 9
706
+ Up on using plane wave expansion,
707
+ ei⃗k·⃗x =
708
+
709
+
710
+ ℓ=0
711
+
712
+
713
+ m=−ℓ
714
+ iℓ jℓ(k x) Yℓm(ˆx) Y ∗
715
+ ℓm(ˆk),
716
+ (3.5)
717
+ and the orthonormal property of spherical harmonics, we obtain
718
+ ⟨aX
719
+ ℓ1m1 aY
720
+ ℓ2m2 aZ
721
+ ℓ3m3 ⟩ = bXYZ
722
+ ℓ1 ℓ2 ℓ3 Gm1 m2 m3
723
+ ℓ1 ℓ2 ℓ3
724
+ ,
725
+ (3.6)
726
+ where all the dependence on m indices are captured in the Gaunt integral
727
+ Gm1 m2 m3
728
+ ℓ1 ℓ2 ℓ3
729
+ =
730
+
731
+ dˆx Yℓ1m1(ˆx) Yℓ2m2(ˆx) Yℓ3m3(ˆx).
732
+ (3.7)
733
+ The quantity bXYZ
734
+ ℓ1 ℓ2 ℓ3 is called the reduced bispectrum and is given by
735
+ bXYZ
736
+ ℓ1 ℓ2 ℓ3 =
737
+ � 2
738
+ π
739
+ �3 �
740
+ x2dx
741
+
742
+ dk1
743
+
744
+ dk2
745
+
746
+ dk3 (k1 k2 k3)2 BR(k1, k2, k3)
747
+ × ∆X
748
+ ℓ1∆Y
749
+ ℓ2∆Z
750
+ ℓ3 jℓ1(k1 x) jℓ2(k2 x) jℓ3(k3 x).
751
+ (3.8)
752
+ The presence of Gaunt integral implies that the reduced bispectra is non-zero only when the
753
+ multipoles satisfies the triangle inequality |ℓ1 − ℓ2| ≤ ℓ3 ≤ |ℓ1 + ℓ2| and when ℓ1 + ℓ2 + ℓ3 is even.
754
+ For isotropic theories, it suffices to work with the reduced bispectrum.
755
+ IV.
756
+ REDUCED BISPECTRA FROM LOOP QUANTUM COSMOLOGY
757
+ We will now compute the reduced bispectrum generated in LQC. The reduced bispectrum
758
+ corresponding to a primordial bispectrum can be computed using Eqn. (3.8). The primordial
759
+ bispectrum corresponding to the non-Gaussianity function Eqn. (2.19) is
760
+ BR(k1, k2, k3) = −6
761
+ 5 (2π2)2
762
+
763
+ f
764
+ bounce
765
+ NL
766
+ e
767
+ −α k1+k2+k3
768
+ kb
769
+ sin
770
+ �k1 + k2 + k3
771
+ kI
772
+
773
+ ×
774
+ �PR(k1)
775
+ k3
776
+ 1
777
+ PR(k2)
778
+ k3
779
+ 2
780
+ + PR(k2)
781
+ k3
782
+ 2
783
+ PR(k3)
784
+ k3
785
+ 3
786
+ + PR(k3)
787
+ k3
788
+ 3
789
+ PR(k1)
790
+ k3
791
+ 1
792
+
793
+ + f
794
+ loc
795
+ NL
796
+ � ¯PR(k1)
797
+ k3
798
+ 1
799
+ ¯PR(k2)
800
+ k3
801
+ 2
802
+ +
803
+ ¯PR(k2)
804
+ k3
805
+ 2
806
+ ¯PR(k3)
807
+ k3
808
+ 3
809
+ +
810
+ ¯PR(k3)
811
+ k3
812
+ 3
813
+ ¯PR(k1)
814
+ k3
815
+ 1
816
+ ��
817
+ .
818
+ (4.1)
819
+ In the above, the bispectrum contains a scale dependent part arising from the bounce and
820
+ a nearly scale invariant part which we have taken to be in the local form. In the latter part,
821
+ we take ¯PR(k) = As (k/k⋆)ns−1. We can substitute Eqn. (4.1) in Eqn. (3.8) to compute the
822
+ reduced bispectrum generated in LQC. However, this calculation involves four integrals, three over
823
+ wavenumbers and one over x variable, which is computationally expensive.
824
+ This calculation can however be simplified, essentially just to two integrals, if we use the separa-
825
+ ble property of the above primordial bispectrum. The total primordial bispectrum is not separable,
826
+ however, the contribution due to the bounce and that due to the scale invariant local template are
827
+ separable. Hence, the reduced bispectrum in LQC can be expressed as
828
+ bXYZ
829
+ ℓ1ℓ2ℓ3 = bbounce
830
+ ℓ1ℓ2ℓ3 + bloc
831
+ ℓ1ℓ2ℓ3,
832
+ (4.2)
833
+
834
+ 10
835
+ where
836
+ bbounce
837
+ ℓ1ℓ2ℓ3
838
+ = −
839
+ � 2
840
+ π
841
+ �3 6
842
+ 5 (2π2)2 f
843
+ bounce
844
+ NL
845
+ � ∞
846
+ 0
847
+ dx x2
848
+
849
+ A
850
+ X
851
+ ℓ1(x) B
852
+ Y
853
+ ℓ2(x) D
854
+ Z
855
+ ℓ3(x) + C
856
+ X
857
+ ℓ1(x) B
858
+ Y
859
+ ℓ2(x)B
860
+ Z
861
+ ℓ3(x)
862
+ +A
863
+ X
864
+ ℓ1(x) D
865
+ Y
866
+ ℓ2(x) B
867
+ Z
868
+ ℓ3(x) + B
869
+ X
870
+ ℓ1(x) A
871
+ Y
872
+ ℓ2(x) D
873
+ Z
874
+ ℓ3(x) + D
875
+ X
876
+ ℓ1(x) A
877
+ Y
878
+ ℓ2(x) B
879
+ Z
880
+ ℓ3(x)
881
+ +B
882
+ X
883
+ ℓ1(x) C
884
+ Y
885
+ ℓ2(x) B
886
+ Z
887
+ ℓ3(x) + B
888
+ X
889
+ ℓ1(x) B
890
+ Y
891
+ ℓ2(x) C
892
+ Z
893
+ ℓ3(x) + D
894
+ X
895
+ ℓ1(x) B
896
+ Y
897
+ ℓ2(x) A
898
+ Z
899
+ ℓ3(x)
900
+ +B
901
+ X
902
+ ℓ1(x) D
903
+ Y
904
+ ℓ2(x) A
905
+ Z
906
+ ℓ3(x) − A
907
+ X
908
+ ℓ1(x) A
909
+ Y
910
+ ℓ2(x) C
911
+ Z
912
+ ℓ3(x) − C
913
+ X
914
+ ℓ1(x) A
915
+ Y
916
+ ℓ2(x) A
917
+ Z
918
+ ℓ3(x)
919
+ −A
920
+ X
921
+ ℓ1(x) C
922
+ Y
923
+ ℓ2(x) A
924
+ Z
925
+ ℓ3(x)
926
+
927
+ (4.3)
928
+ and
929
+ bloc
930
+ ℓ1ℓ2ℓ3 = −
931
+ � 2
932
+ π
933
+ �3 6
934
+ 5 (2π2)2 f
935
+ loc
936
+ NL
937
+ � ∞
938
+ 0
939
+ dx x2
940
+
941
+ E
942
+ X
943
+ ℓ1(x) E
944
+ Y
945
+ ℓ2(x) G
946
+ Z
947
+ ℓ3(x) + G
948
+ X
949
+ ℓ1(x) E
950
+ Y
951
+ ℓ2(x) E
952
+ Z
953
+ ℓ3(x)
954
+ + E
955
+ X
956
+ ℓ1(x) G
957
+ Y
958
+ ℓ2(x) E
959
+ Z
960
+ ℓ3(x)
961
+
962
+ .
963
+ (4.4)
964
+ In the above expressions, the functions A
965
+ X
966
+ ℓ (x), B
967
+ X
968
+ ℓ (x), C
969
+ X
970
+ ℓ (x), D
971
+ X
972
+ ℓ (x), E
973
+ X
974
+ ℓ (x) and G
975
+ X
976
+ ℓ (x) are
977
+ A
978
+ X
979
+ ℓ (x) =
980
+ � ∞
981
+ 0
982
+ dk ∆X
983
+ ℓ (k) jℓ(kx) PR(k)
984
+ k
985
+ e
986
+ − αk
987
+ kb sin( k
988
+ kI
989
+ ),
990
+ (4.5a)
991
+ B
992
+ X
993
+ ℓ (x) =
994
+ � ∞
995
+ 0
996
+ dk ∆X
997
+ ℓ (k) jℓ(kx) PR(k)
998
+ k
999
+ e
1000
+ − αk
1001
+ kb cos( k
1002
+ kI
1003
+ ),
1004
+ (4.5b)
1005
+ C
1006
+ X
1007
+ ℓ (x) =
1008
+ � ∞
1009
+ 0
1010
+ dk ∆X
1011
+ ℓ (k) jℓ(kx) k2 e
1012
+ − αk
1013
+ kb sin( k
1014
+ kI
1015
+ ),
1016
+ (4.5c)
1017
+ D
1018
+ X
1019
+ ℓ (x) =
1020
+ � ∞
1021
+ 0
1022
+ dk ∆X
1023
+ ℓ (k) jl(kx) k2e
1024
+ − αk
1025
+ kb cos( k
1026
+ kI
1027
+ ),
1028
+ (4.5d)
1029
+ E
1030
+ X
1031
+ ℓ (x) =
1032
+ � ∞
1033
+ 0
1034
+ dk ∆X
1035
+ ℓ (k) jℓ(kx) k−1As(k/k∗)ns−1,
1036
+ (4.5e)
1037
+ G
1038
+ X
1039
+ ℓ (x) =
1040
+ � ∞
1041
+ 0
1042
+ dk ∆X
1043
+ ℓ (k) jℓ(kx) k2.
1044
+ (4.5f)
1045
+ Note that, each of the functions A
1046
+ X
1047
+ ℓ (x), B
1048
+ X
1049
+ ℓ (x), C
1050
+ X
1051
+ ℓ (x), D
1052
+ X
1053
+ ℓ (x), E
1054
+ X
1055
+ ℓ (x) and G
1056
+ X
1057
+ ℓ (x) involve an
1058
+ integral over the wavenumber. The reduced bispectrum can now be calculated by evaluating these
1059
+ functions for all the required values of multipoles and then finally performing the integrals Eqns.
1060
+ (4.3, 4.4).
1061
+ V.
1062
+ NUMERICAL PROCEDURE AND RESULTS
1063
+ We now discuss the numerical procedure we have followed for computing the reduced bispectrum
1064
+ generated in LQC. The first step in calculating reduced bispectrum is the evaluation of functions
1065
+ Eqns. (4.5). In order to compute these functions, we require the transfer functions ∆X
1066
+ ℓ , where X
1067
+ can be either temperature fluctuations or electric polarisation. We use publicly available Boltzmann
1068
+ code class [69] to generate both the transfer functions. We perform the integral using Simpson’s
1069
+ rule. We choose this method since the integrand is highly oscillatory and this gives better accuracy
1070
+ when we work with sufficiently small step size.
1071
+ Since the scale of oscillations occur at kI =
1072
+ 10−7 Mpc−1, we have worked with a step size of ∆k = 10−8 Mpc−1. This leads to an accuracy
1073
+ of O(10−32). The behaviour of functions A
1074
+ X
1075
+ ℓ (x), B
1076
+ X
1077
+ ℓ (x), C
1078
+ X
1079
+ ℓ (x), D
1080
+ X
1081
+ ℓ (x), E
1082
+ X
1083
+ ℓ (x) and G
1084
+ X
1085
+ ℓ (x) for
1086
+
1087
+ 11
1088
+ 102
1089
+ 103
1090
+ 104
1091
+ x (MPc)
1092
+ 10−29
1093
+ 10−26
1094
+ 10−23
1095
+ 10−20
1096
+ 10−17
1097
+ 10−14
1098
+ 10−11
1099
+ 10−8
1100
+ ���A
1101
+ T
1102
+ 4(x)
1103
+ ���
1104
+ ���B
1105
+ T
1106
+ 4(x)
1107
+ ���
1108
+ ���C
1109
+ T
1110
+ 4 (x)
1111
+ ���
1112
+ ���D
1113
+ T
1114
+ 4(x)
1115
+ ���
1116
+ ���E
1117
+ T
1118
+ 4(x)
1119
+ ���
1120
+ ���G
1121
+ T
1122
+ 4(x)
1123
+ ���
1124
+ 102
1125
+ 103
1126
+ 104
1127
+ x (MPc)
1128
+ 10−29
1129
+ 10−26
1130
+ 10−23
1131
+ 10−20
1132
+ 10−17
1133
+ 10−14
1134
+ ���A
1135
+ E
1136
+ 4(x)
1137
+ ���
1138
+ ���B
1139
+ E
1140
+ 4 (x)
1141
+ ���
1142
+ ���C
1143
+ E
1144
+ 4 (x)
1145
+ ���
1146
+ ���D
1147
+ E
1148
+ 4(x)
1149
+ ���
1150
+ ���E
1151
+ E
1152
+ 4 (x)
1153
+ ���
1154
+ ���G
1155
+ E
1156
+ 4(x)
1157
+ ���
1158
+ 102
1159
+ 103
1160
+ 104
1161
+ x (MPc)
1162
+ 10−97
1163
+ 10−85
1164
+ 10−73
1165
+ 10−61
1166
+ 10−49
1167
+ 10−37
1168
+ 10−25
1169
+ 10−13
1170
+ ���A
1171
+ T
1172
+ 40(x)
1173
+ ���
1174
+ ���B
1175
+ T
1176
+ 40(x)
1177
+ ���
1178
+ ���C
1179
+ T
1180
+ 40(x)
1181
+ ���
1182
+ ���D
1183
+ T
1184
+ 40(x)
1185
+ ���
1186
+ ���E
1187
+ T
1188
+ 40(x)
1189
+ ���
1190
+ ���G
1191
+ T
1192
+ 40(x)
1193
+ ���
1194
+ 102
1195
+ 103
1196
+ 104
1197
+ x (MPc)
1198
+ 10−97
1199
+ 10−85
1200
+ 10−73
1201
+ 10−61
1202
+ 10−49
1203
+ 10−37
1204
+ 10−25
1205
+ 10−13
1206
+ ���A
1207
+ E
1208
+ 40(x)
1209
+ ���
1210
+ ���B
1211
+ E
1212
+ 40(x)
1213
+ ���
1214
+ ���C
1215
+ E
1216
+ 40(x)
1217
+ ���
1218
+ ���D
1219
+ E
1220
+ 40(x)
1221
+ ���
1222
+ ���E
1223
+ E
1224
+ 40(x)
1225
+ ���
1226
+ ���G
1227
+ E
1228
+ 40(x)
1229
+ ���
1230
+ FIG. 4. The behaviour of functions in Eqn. (4.5) with x for multipoles ℓ = 4 and 40. The contribution from
1231
+ the local part of the bispectrum viz.
1232
+ E
1233
+ X
1234
+ ℓ (x) and G
1235
+ X
1236
+ ℓ (x) are clearly dominant compared to those arising
1237
+ from the bounce part viz.
1238
+ A
1239
+ X
1240
+ ℓ (x), B
1241
+ X
1242
+ ℓ (x), C
1243
+ X
1244
+ ℓ (x) and D
1245
+ X
1246
+ ℓ (x).
1247
+ multipoles ℓ = 4 and 40 are shown in figure 4. From the figure, it is clear that the functions
1248
+ E
1249
+ X
1250
+ ℓ (x) and G
1251
+ X
1252
+ ℓ (x) are dominant compared to A
1253
+ X
1254
+ ℓ (x), B
1255
+ X
1256
+ ℓ (x), C
1257
+ X
1258
+ ℓ (x) and D
1259
+ X
1260
+ ℓ (x).
1261
+ This is an
1262
+ indication of the fact that local part of the bispectrum is dominant compared to the oscillatory
1263
+ part. The next step in computing reduced bispectrum is the evaluation of Eqns. (4.3, 4.4). We
1264
+ perform these integrals over x with a step size of 50 in the range x ∈ [0, 40000]. We have made the
1265
+ calculations faster by using vectorization available in NumPy and by parallelizing the computation
1266
+ wherever possible.
1267
+ Reduced bispectra bTTT
1268
+ ℓ1, ℓ2, ℓ3, bTTE
1269
+ ℓ1, ℓ2, ℓ3, bTEE
1270
+ ℓ1, ℓ2, ℓ3 and bEEE
1271
+ ℓ1, ℓ2, ℓ3 generated in LQC are shown in figures
1272
+ 5, 6, 7 and 8 respectively. We have illustrated two different configurations of the bispectra. In
1273
+ these figures, we have separately plotted the contribution from the local and bounce parts of the
1274
+ bispectrum. The figures show that the contribution to the bispectrum from the oscillatory part
1275
+ of the template is negligible compared to that from the local part. This shows that the reduced
1276
+ bispectra generated in LQC will be similar to that produced in slow roll inflation and hence will
1277
+ be consistent with observations by Planck [10].
1278
+
1279
+ 12
1280
+ 0
1281
+ 10
1282
+ 20
1283
+ 30
1284
+ 40
1285
+ 50
1286
+
1287
+ 10−60
1288
+ 10−55
1289
+ 10−50
1290
+ 10−45
1291
+ 10−40
1292
+ 10−35
1293
+ 10−30
1294
+ 10−25
1295
+ 10−20
1296
+ ���bTTT
1297
+ ℓ,ℓ,ℓ
1298
+ ���
1299
+ local
1300
+ bounce
1301
+ 0
1302
+ 10
1303
+ 20
1304
+ 30
1305
+ 40
1306
+ 50
1307
+
1308
+ 10−50
1309
+ 10−46
1310
+ 10−42
1311
+ 10−38
1312
+ 10−34
1313
+ 10−30
1314
+ 10−26
1315
+ 10−22
1316
+ ���bTTT
1317
+ 2,ℓ,ℓ
1318
+ ���
1319
+ local
1320
+ bounce
1321
+ FIG. 5.
1322
+ The reduced bispectra bTTT
1323
+ ℓ1, ℓ2, ℓ3 in two different configurations. The plots illustrate that bloc
1324
+ ℓ1ℓ2ℓ3 >>
1325
+ bbounce
1326
+ ℓ1ℓ2ℓ3 .
1327
+ 0
1328
+ 10
1329
+ 20
1330
+ 30
1331
+ 40
1332
+ 50
1333
+
1334
+ 10−54
1335
+ 10−46
1336
+ 10−38
1337
+ 10−30
1338
+ 10−23
1339
+ ���bTTE
1340
+ ℓ,ℓ,ℓ
1341
+ ���
1342
+ local
1343
+ bounce
1344
+ 0
1345
+ 10
1346
+ 20
1347
+ 30
1348
+ 40
1349
+ 50
1350
+
1351
+ 10−51
1352
+ 10−44
1353
+ 10−37
1354
+ 10−30
1355
+ 10−23
1356
+ ���bTTE
1357
+ 2,ℓ,ℓ
1358
+ ���
1359
+ local
1360
+ bounce
1361
+ FIG. 6. The plots of reduced bispectra bTTE
1362
+ ℓ1, ℓ2, ℓ3 in two different configurations. Note that that bloc
1363
+ ℓ1ℓ2ℓ3 >>
1364
+ bbounce
1365
+ ℓ1ℓ2ℓ3 .
1366
+ 0
1367
+ 10
1368
+ 20
1369
+ 30
1370
+ 40
1371
+ 50
1372
+
1373
+ 10−57
1374
+ 10−49
1375
+ 10−41
1376
+ 10−33
1377
+ 10−25
1378
+ ���bTEE
1379
+ ℓ,ℓ,ℓ
1380
+ ���
1381
+ local
1382
+ bounce
1383
+ 0
1384
+ 10
1385
+ 20
1386
+ 30
1387
+ 40
1388
+ 50
1389
+
1390
+ 10−50
1391
+ 10−43
1392
+ 10−37
1393
+ 10−31
1394
+ 10−25
1395
+ ���bTEE
1396
+ 2,ℓ,ℓ
1397
+ ���
1398
+ local
1399
+ bounce
1400
+ FIG. 7. The plots of reduced bispectra bTEE
1401
+ ℓ1, ℓ2, ℓ3 in two different configurations. Clearly, the bloc
1402
+ ℓ1ℓ2ℓ3 is much
1403
+ larger than bbounce
1404
+ ℓ1ℓ2ℓ3 .
1405
+
1406
+ 13
1407
+ 0
1408
+ 10
1409
+ 20
1410
+ 30
1411
+ 40
1412
+ 50
1413
+
1414
+ 10−55
1415
+ 10−48
1416
+ 10−41
1417
+ 10−34
1418
+ 10−27
1419
+ ���bEEE
1420
+ ℓ,ℓ,ℓ
1421
+ ���
1422
+ local
1423
+ bounce
1424
+ 0
1425
+ 10
1426
+ 20
1427
+ 30
1428
+ 40
1429
+ 50
1430
+
1431
+ 10−55
1432
+ 10−48
1433
+ 10−41
1434
+ 10−34
1435
+ 10−27
1436
+ ���bEEE
1437
+ 2,ℓ,ℓ
1438
+ ���
1439
+ local
1440
+ bounce
1441
+ FIG. 8.
1442
+ The plots of reduced bispectra bEEE
1443
+ ℓ1, ℓ2, ℓ3 in two different configurations. Note that, the reduced
1444
+ bispectrum is dominated by contribution from the local part of the template.
1445
+ VI.
1446
+ SUMMARY AND DISCUSSION
1447
+ State of the art measurements of CMB by Planck has put strong constraints on primordial
1448
+ non-Gaussianity [10]. Observations by Planck point towards a small primordial non-Gaussianity
1449
+ which is consistent with the one generated in slow roll models of inflation. In LQC, primordial
1450
+ perturbations originate in an adiabatic vacuum before the bounce. These then evolve through
1451
+ the bounce, then through the inflationary epoch before their amplitude freezes upon horizon exit
1452
+ during inflation.
1453
+ The quantum bounce sets a scale kLQC in the problem.
1454
+ Modes which have
1455
+ wavenumbers comparable to or smaller than kLQC are excited during the bounce and modes with
1456
+ larger wavenumbers are not.
1457
+ This implies that modes with k ≲ kLQC are in an excited and
1458
+ non-Gaussian state at the onset of inflation. This non-Gaussianity is then further enhanced as
1459
+ the modes exit the horizon during inflation.
1460
+ However, modes with longer wavenumbers, since
1461
+ they are not excited during the bounce, behave very similarly to modes in slow roll inflation.
1462
+ Hence, we have a situation where longer wavelength modes are non-Gaussian where as the shorter
1463
+ ones remain Gaussian. In order to establish the viability of LQC as a model for pre-inflationary
1464
+ universe, it is important to answer whether LQC is compatible with the constraints on primordial
1465
+ non-Gaussianity set by Planck.
1466
+ With this goal, we investigated the imprints of primordial non-Gaussianity in the bispectrum of
1467
+ temperature and electric polarisation and their cross-correlations generated in LQC. In particular,
1468
+ motivated by previous efforts, we proposed a template which captures the essential features of
1469
+ the primordial bispectrum generated in LQC. We then used the template to compute the bTTT
1470
+ ℓ1 ℓ2 ℓ3,
1471
+ bTTE
1472
+ ℓ1 ℓ2 ℓ3, bTEE
1473
+ ℓ1 ℓ2 ℓ3 and bEEE
1474
+ ℓ1 ℓ2 ℓ3 bispectra. To simplify the calculation, we used the separable property
1475
+ of the proposed template. We considered the bispectra in LQC as consisting of two terms, one
1476
+ scale dependent and oscillatory part arising from the bounce and the other nearly scale invariant
1477
+ part arising from slow roll inflation. We find that, the contribution from the bounce to the reduced
1478
+ bispectra is negligible compared to that arising from the part corresponding to slow roll inflation.
1479
+ This implies that the reduced bispectrum generated in LQC is similar to that generated in slow roll
1480
+ inflation. Hence, we conclude that the primordial non-Gaussianity generated in LQC is compatible
1481
+ with the constraints from Planck. This is the central result of this paper.
1482
+ The primordial perturbations generated in LQC is non-Gaussian in nature, yet our computation
1483
+ illustrates that the reduced bispectra of temperature and electric polarisation are similar to that of
1484
+ slow roll. This seemingly contradicting finding is because of the oscillatory nature of the primordial
1485
+
1486
+ 14
1487
+ bispectrum. The reduced bispectra involves integrals over wavenumbers which average over these
1488
+ oscillations. Our result could be compared with those of [70, 71] where they had worked with a
1489
+ non-oscillatory template. While they found that, in the absence of oscillations, the contribution
1490
+ from the bounce is significant enough to be observed by Planck, we find that presence of oscillations
1491
+ in the primordial bispectra dilutes any imprints of non-Gaussianity on the reduced bispectra. Thus,
1492
+ it should be highlighted that a small reduced bispectra need not necessarily imply the absence of
1493
+ primordial non-Gaussianity. Hence, it would also be interesting to look for any other measurable
1494
+ imprints of such oscillatory and scale dependent primordial non-Gaussianity.
1495
+ Finally, our findings are relevant for constraints on the amount of pre-inflationary expansion in
1496
+ LQC. In LQC, the amount of expansion before inflationary epoch is set by the value of scalar field
1497
+ at the bounce. The scalar field rolls up the potential after the bounce, comes to rest momentarily
1498
+ before it rolls down and settles in to the inflationary attractor. Hence, the value of the scalar field
1499
+ at the bounce determines the amount of expansion between the bounce and the onset of inflation.
1500
+ This epoch of expansion is relevant as it determines whether the scales that are sensitive to the
1501
+ effects of the bounce are visible today. If this epoch of pre-inflationary expansion is very large, then
1502
+ the imprints of the bounce will not be visible in the Universe today. However, if the pre-inflationary
1503
+ expansion is small, then the primordial power spectrum and bispectrum will be scale dependent at
1504
+ observable scales. The power spectrum of temperature fluctuations fits extremely well to those due
1505
+ to a nearly scale invariant primordial power spectrum at multipoles ℓ > 30 [1, 51]. This imposes
1506
+ a lower limit to the amount of pre-inflationary expansion and hence a lower limit to the value of
1507
+ scalar field at the bounce [29].
1508
+ Compared to the primordial power spectrum, the primordial non-Gaussianity is more sensitive
1509
+ to the bounce, i.e.
1510
+ fNL(k1, k2, k3) is scale dependent at larger wavenumbers than the primordial
1511
+ power spectrum. This leads to a question whether imprints of primordial non-Gaussianity leads
1512
+ to a stronger lower limit to the pre-inflationary expansion. Our calculations, carried out in this
1513
+ work, answer this question in the negative. More specifically, since the reduced bispectra do not
1514
+ carry any imprints of the bounce, we find that it do not provide any constraints on the epoch of
1515
+ pre-inflationary expansion.
1516
+ ACKNOWLEDGEMENT
1517
+ We thank Ivan Agullo for his comments. This work was supported by Science and Engineering
1518
+ Research Board (SERB) through Start-up Research Grant SRG/2021/001769. We acknowledge
1519
+ the use of PU HPC facility of the National Supercomputing Mission project.
1520
+ [1] N.
1521
+ Aghanim
1522
+ et
1523
+ al.
1524
+ (Planck),
1525
+ Planck
1526
+ 2018
1527
+ results.
1528
+ VI.
1529
+ Cosmological
1530
+ parameters,
1531
+ (2018),
1532
+ arXiv:1807.06209 [astro-ph.CO].
1533
+ [2] A. Riotto, Inflation and the theory of cosmological perturbations, ICTP Lect. Notes Ser. 14, 317 (2003),
1534
+ arXiv:hep-ph/0210162.
1535
+ [3] B. A. Bassett, S. Tsujikawa, and D. Wands, Inflation dynamics and reheating, Rev. Mod. Phys. 78,
1536
+ 537 (2006), arXiv:astro-ph/0507632.
1537
+ [4] L. Sriramkumar, An introduction to inflation and cosmological perturbation theory,
1538
+ (2009),
1539
+ arXiv:0904.4584 [astro-ph.CO].
1540
+ [5] D. Baumann, Inflation, in Theoretical Advanced Study Institute in Elementary Particle Physics: Physics
1541
+ of the Large and the Small (2011) pp. 523–686, arXiv:0907.5424 [hep-th].
1542
+ [6] Y. Akrami et al. (Planck), Planck 2018 results. X. Constraints on inflation, (2018), arXiv:1807.06211
1543
+ [astro-ph.CO].
1544
+
1545
+ 15
1546
+ [7] J. Martin, C. Ringeval, R. Trotta, and V. Vennin, The Best Inflationary Models After Planck, JCAP
1547
+ 03, 039, arXiv:1312.3529 [astro-ph.CO].
1548
+ [8] A. Ijjas, P. J. Steinhardt, and A. Loeb, Cosmic inflation theory faces challenges, Scientific American
1549
+ 316 (2017).
1550
+ [9] A. Guth et al., A cosmic controversy, Scientific American 316 (2017).
1551
+ [10] Y. Akrami et al. (Planck), Planck 2018 results. IX. Constraints on primordial non-Gaussianity, Astron.
1552
+ Astrophys. 641, A9 (2020), arXiv:1905.05697 [astro-ph.CO].
1553
+ [11] A. Ashtekar, T. Pawlowski, and P. Singh, Quantum nature of the big bang, Phys. Rev. Lett. 96, 141301
1554
+ (2006), arXiv:gr-qc/0602086 [gr-qc].
1555
+ [12] A. Ashtekar, T. Pawlowski, and P. Singh, Quantum Nature of the Big Bang: Improved dynamics, Phys.
1556
+ Rev. D74, 084003 (2006), arXiv:gr-qc/0607039 [gr-qc].
1557
+ [13] A. Ashtekar and P. Singh, Loop Quantum Cosmology: A Status Report, Class. Quant. Grav. 28, 213001
1558
+ (2011), arXiv:1108.0893 [gr-qc].
1559
+ [14] I. Agullo and P. Singh, Loop Quantum Cosmology, in Loop Quantum Gravity: The First 30 Years,
1560
+ edited by A. Ashtekar and J. Pullin (WSP, 2017) pp. 183–240, arXiv:1612.01236 [gr-qc].
1561
+ [15] I. Agullo and A. Corichi, Loop Quantum Cosmology, in Springer Handbook of Spacetime, edited by
1562
+ A. Ashtekar and V. Petkov (2014) pp. 809–839, arXiv:1302.3833 [gr-qc].
1563
+ [16] M. Bojowald, Absence of singularity in loop quantum cosmology, Phys. Rev. Lett. 86, 5227 (2001),
1564
+ arXiv:gr-qc/0102069.
1565
+ [17] A. Ashtekar, M. Bojowald, and J. Lewandowski, Mathematical structure of loop quantum cosmology,
1566
+ Adv. Theor. Math. Phys. 7, 233 (2003), arXiv:gr-qc/0304074.
1567
+ [18] G. A. Mena Marugan, Loop Quantum Cosmology: A cosmological theory with a view, Proceedings,
1568
+ Spanish Relativity Meeting: Gravity as a crossroad in physics (ERE 2010): Granada, Spain, September
1569
+ 6-10, 2010, J. Phys. Conf. Ser. 314, 012012 (2011), arXiv:1101.1738 [gr-qc].
1570
+ [19] K. Banerjee, G. Calcagni, and M. Martin-Benito, Introduction to loop quantum cosmology, SIGMA 8,
1571
+ 016 (2012), arXiv:1109.6801 [gr-qc].
1572
+ [20] M. Bojowald, G. M. Hossain, M. Kagan, and S. Shankaranarayanan, Gauge invariant cosmological
1573
+ perturbation equations with corrections from loop quantum gravity, Phys. Rev. D 79, 043505 (2009),
1574
+ [Erratum: Phys.Rev.D 82, 109903 (2010)], arXiv:0811.1572 [gr-qc].
1575
+ [21] M. Bojowald and G. Calcagni, Inflationary observables in loop quantum cosmology, JCAP 03, 032,
1576
+ arXiv:1011.2779 [gr-qc].
1577
+ [22] I. Agullo, A. Ashtekar, and W. Nelson, A Quantum Gravity Extension of the Inflationary Scenario,
1578
+ Phys. Rev. Lett. 109, 251301 (2012), arXiv:1209.1609 [gr-qc].
1579
+ [23] I. Agullo, A. Ashtekar, and W. Nelson, Extension of the quantum theory of cosmological perturbations
1580
+ to the Planck era, Phys. Rev. D 87, 043507 (2013), arXiv:1211.1354 [gr-qc].
1581
+ [24] I. Agullo, A. Ashtekar, and W. Nelson, The pre-inflationary dynamics of loop quantum cosmology: Con-
1582
+ fronting quantum gravity with observations, Class. Quant. Grav. 30, 085014 (2013), arXiv:1302.0254
1583
+ [gr-qc].
1584
+ [25] M. Fern´andez-M´endez, G. A. Mena Marug´an, and J. Olmedo, Hybrid quantization of an inflationary
1585
+ model: The flat case, Phys. Rev. D 88, 044013 (2013), arXiv:1307.5222 [gr-qc].
1586
+ [26] M. Fern´andez-M´endez, G. A. Mena Marug´an, and J. Olmedo, Effective dynamics of scalar perturbations
1587
+ in a flat Friedmann-Robertson-Walker spacetime in Loop Quantum Cosmology, Phys. Rev. D 89,
1588
+ 044041 (2014), arXiv:1401.5256 [gr-qc].
1589
+ [27] A. Barrau, M. Bojowald, G. Calcagni, J. Grain, and M. Kagan, Anomaly-free cosmological perturba-
1590
+ tions in effective canonical quantum gravity, JCAP 05, 051, arXiv:1404.1018 [gr-qc].
1591
+ [28] D. M. de Blas and J. Olmedo, Primordial power spectra for scalar perturbations in loop quantum
1592
+ cosmology, JCAP 1606 (06), 029, arXiv:1601.01716 [gr-qc].
1593
+ [29] I. Agullo and N. A. Morris, Detailed analysis of the predictions of loop quantum cosmology for the
1594
+ primordial power spectra, Phys. Rev. D92, 124040 (2015), arXiv:1509.05693 [gr-qc].
1595
+ [30] I. Agullo, A. Ashtekar, and B. Gupt, Phenomenology with fluctuating quantum geometries in loop
1596
+ quantum cosmology, Class. Quant. Grav. 34, 074003 (2017), arXiv:1611.09810 [gr-qc].
1597
+ [31] A. Ashtekar and B. Gupt, Quantum Gravity in the Sky: Interplay between fundamental theory and
1598
+ observations, Class. Quant. Grav. 34, 014002 (2017), arXiv:1608.04228 [gr-qc].
1599
+ [32] F. B. Mart´ınez and J. Olmedo, Primordial tensor modes of the early Universe, Phys. Rev. D 93, 124008
1600
+ (2016), arXiv:1605.04293 [gr-qc].
1601
+
1602
+ 16
1603
+ [33] L. Castell´o Gomar, G. A. Mena Marug´an, D. Mart´ın De Blas, and J. Olmedo, Hybrid loop quantum
1604
+ cosmology and predictions for the cosmic microwave background, Phys. Rev. D 96, 103528 (2017),
1605
+ arXiv:1702.06036 [gr-qc].
1606
+ [34] T. Zhu, A. Wang, G. Cleaver, K. Kirsten, and Q. Sheng, Pre-inflationary universe in loop quantum
1607
+ cosmology, Phys. Rev. D96, 083520 (2017), arXiv:1705.07544 [gr-qc].
1608
+ [35] I. Agullo, Primordial power spectrum from the Dapor-Liegener model of loop quantum cosmology, Gen.
1609
+ Rel. Grav. 50, 91 (2018), arXiv:1805.11356 [gr-qc].
1610
+ [36] B.-F. Li, P. Singh, and A. Wang, Primordial power spectrum from the dressed metric approach in loop
1611
+ cosmologies, Phys. Rev. D 101, 086004 (2020), arXiv:1912.08225 [gr-qc].
1612
+ [37] B. Elizaga Navascu´es, G. A. M. Marug´an, and S. Prado, Non-oscillating power spectra in Loop Quantum
1613
+ Cosmology, Class. Quant. Grav. 38, 035001 (2020), arXiv:2005.10194 [gr-qc].
1614
+ [38] B.-F. Li, J. Olmedo, P. Singh, and A. Wang, Primordial scalar power spectrum from the hybrid approach
1615
+ in loop cosmologies, Phys. Rev. D 102, 126025 (2020), arXiv:2008.09135 [gr-qc].
1616
+ [39] I. Agullo, J. Olmedo, and V. Sreenath, Predictions for the Cosmic Microwave Background from an
1617
+ Anisotropic Quantum Bounce, Phys. Rev. Lett. 124, 251301 (2020), arXiv:2003.02304 [gr-qc].
1618
+ [40] I. Agullo, J. Olmedo, and V. Sreenath, Observational consequences of Bianchi I spacetimes in loop
1619
+ quantum cosmology, Phys. Rev. D 102, 043523 (2020), arXiv:2006.01883 [gr-qc].
1620
+ [41] A. Ashtekar, B. Gupt, D. Jeong, and V. Sreenath, Alleviating the Tension in the Cosmic Microwave
1621
+ Background using Planck-Scale Physics, Phys. Rev. Lett. 125, 051302 (2020), arXiv:2001.11689 [astro-
1622
+ ph.CO].
1623
+ [42] B. Elizaga Navascu´es and G. A. M. Marug´an, Hybrid Loop Quantum Cosmology: An Overview, (2020),
1624
+ arXiv:2011.04559 [gr-qc].
1625
+ [43] M. Mart´ın-Benito, R. B. Neves, and J. Olmedo, States of Low Energy in bouncing inflationary scenarios
1626
+ in Loop Quantum Cosmology, (2021), arXiv:2104.03035 [gr-qc].
1627
+ [44] B.-F. Li, P. Singh, and A. Wang, Phenomenological implications of modified loop cosmologies: an
1628
+ overview, Front. Astron. Space Sci. 8, 701417 (2021), arXiv:2105.14067 [gr-qc].
1629
+ [45] A. Ashtekar, B. Gupt, and V. Sreenath, Cosmic Tango Between the Very Small and the Very Large:
1630
+ Addressing CMB Anomalies Through Loop Quantum Cosmology, Front. Astron. Space Sci. 8, 76
1631
+ (2021), arXiv:2103.14568 [gr-qc].
1632
+ [46] I. Agullo, D. Kranas, and V. Sreenath, Anomalies in the Cosmic Microwave Background and Their
1633
+ Non-Gaussian Origin in Loop Quantum Cosmology, Front. Astron. Space Sci. 8, 703845 (2021),
1634
+ arXiv:2105.12993 [gr-qc].
1635
+ [47] I. Agullo, B. Bolliet, and V. Sreenath, Non-Gaussianity in Loop Quantum Cosmology, Phys. Rev. D97,
1636
+ 066021 (2018), arXiv:1712.08148 [gr-qc].
1637
+ [48] I. Agullo, Loop quantum cosmology, non-Gaussianity, and CMB power asymmetry, Phys. Rev. D92,
1638
+ 064038 (2015), arXiv:1507.04703 [gr-qc].
1639
+ [49] V. Sreenath, I. Agullo, and B. Bolliet, Computation of non-Gaussianity in loop quantum cosmology,
1640
+ in 15th Marcel Grossmann Meeting on Recent Developments in Theoretical and Experimental General
1641
+ Relativity, Astrophysics, and Relativistic Field Theories (2019) arXiv:1904.01075 [gr-qc].
1642
+ [50] T. Zhu, A. Wang, K. Kirsten, G. Cleaver, and Q. Sheng, Primordial non-Gaussianity and power asym-
1643
+ metry with quantum gravitational effects in loop quantum cosmology, Phys. Rev. D 97, 043501 (2018),
1644
+ arXiv:1709.07479 [astro-ph.CO].
1645
+ [51] N. Aghanim et al. (Planck), Planck 2018 results. V. CMB power spectra and likelihoods,
1646
+ (2019),
1647
+ arXiv:1907.12875 [astro-ph.CO].
1648
+ [52] P. Diener, B. Gupt, and P. Singh, Chimera: A hybrid approach to numerical loop quantum cosmology,
1649
+ Class. Quant. Grav. 31, 025013 (2014), arXiv:1310.4795 [gr-qc].
1650
+ [53] P. Diener, B. Gupt, and P. Singh, Numerical simulations of a loop quantum cosmos: robustness of
1651
+ the quantum bounce and the validity of effective dynamics, Class. Quant. Grav. 31, 105015 (2014),
1652
+ arXiv:1402.6613 [gr-qc].
1653
+ [54] A. Ashtekar and D. Sloan, Loop quantum cosmology and slow roll inflation, Phys. Lett. B694, 108
1654
+ (2011), arXiv:0912.4093 [gr-qc].
1655
+ [55] A. Ashtekar and D. Sloan, Probability of Inflation in Loop Quantum Cosmology, Gen. Rel. Grav. 43,
1656
+ 3619 (2011), arXiv:1103.2475 [gr-qc].
1657
+ [56] B. Bolliet, A. Barrau, K. Martineau, and F. Moulin, Some Clarifications on the Duration of Inflation
1658
+ in Loop Quantum Cosmology, Class. Quant. Grav. 34, 145003 (2017), arXiv:1701.02282 [gr-qc].
1659
+
1660
+ 17
1661
+ [57] B. Bonga and B. Gupt, Inflation with the Starobinsky potential in Loop Quantum Cosmology, Gen.
1662
+ Rel. Grav. 48, 71 (2016), arXiv:1510.00680 [gr-qc].
1663
+ [58] B. Bonga and B. Gupt, Phenomenological investigation of a quantum gravity extension of inflation
1664
+ with the Starobinsky potential, Phys. Rev. D93, 063513 (2016), arXiv:1510.04896 [gr-qc].
1665
+ [59] J. M. Maldacena, Non-Gaussian features of primordial fluctuations in single field inflationary models,
1666
+ JHEP 05, 013, arXiv:astro-ph/0210603 [astro-ph].
1667
+ [60] I. Agullo, D. Kranas, and V. Sreenath, Anomalies in the CMB from a cosmic bounce, Gen. Rel. Grav.
1668
+ 53, 17 (2021), arXiv:2005.01796 [astro-ph.CO].
1669
+ [61] I. Agullo, D. Kranas, and V. Sreenath, Large scale anomalies in the CMB and non-Gaussianity in
1670
+ bouncing cosmologies, Class. Quant. Grav. 38, 065010 (2021), arXiv:2006.09605 [astro-ph.CO].
1671
+ [62] R. Durrer, The Cosmic Microwave Background, 2nd ed. (Cambridge University Press, 2020).
1672
+ [63] S. Dodelson, Modern Cosmology (Academic Press, Amsterdam, 2003).
1673
+ [64] S. Weinberg, Cosmology (Oxford University Press, 2008).
1674
+ [65] E. Komatsu and D. N. Spergel, Acoustic signatures in the primary microwave background bispectrum,
1675
+ Phys. Rev. D 63, 063002 (2001), arXiv:astro-ph/0005036.
1676
+ [66] J. R. Fergusson and E. P. S. Shellard, Primordial non-Gaussianity and the CMB bispectrum, Phys.
1677
+ Rev. D 76, 083523 (2007), arXiv:astro-ph/0612713.
1678
+ [67] M. Liguori, E. Sefusatti, J. R. Fergusson, and E. P. S. Shellard, Primordial Non-Gaussianity and
1679
+ Bispectrum Measurements in the Cosmic Microwave Background and Large-Scale Structure, Advances
1680
+ in Astronomy 2010, 980523 (2010), arXiv:1001.4707 [astro-ph.CO].
1681
+ [68] J. R. Fergusson, M. Liguori, and E. P. S. Shellard, The CMB bispectrum, JCAP 2012 (12), 032,
1682
+ arXiv:1006.1642 [astro-ph.CO].
1683
+ [69] D. Blas, J. Lesgourgues, and T. Tram, The Cosmic Linear Anisotropy Solving System (CLASS). Part
1684
+ II: Approximation schemes, JCAP 7, 034, arXiv:1104.2933.
1685
+ [70] P. C. M. Delgado, R. Durrer, and N. Pinto-Neto, The CMB bispectrum from bouncing cosmologies,
1686
+ JCAP 11, 024, arXiv:2108.06175 [astro-ph.CO].
1687
+ [71] B. van Tent, P. C. M. Delgado, and R. Durrer, Constraining the bispectrum from bouncing cosmologies
1688
+ with Planck, (2022), arXiv:2212.05977 [astro-ph.CO].
1689
+
Z9E5T4oBgHgl3EQfDg58/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
ZNAyT4oBgHgl3EQfifg0/content/tmp_files/2301.00395v1.pdf.txt ADDED
@@ -0,0 +1,2656 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ CORGI-PM
2
+ : A Chinese Corpus For Gender Bias Probing and
3
+ Mitigation
4
+ Ge Zhang1 3 4 ∗, Yizhi Li2 ∗, Yaoyao Wu5, Linyuan Zhang 6, Chenghua Lin 2 †, Jiayi Geng7, Shi Wang 3 †, Jie Fu 1
5
+ 1 Beijing Academy of Artificial Intelligence, China
6
+ 2 Department of Computer Science, The University of Sheffield, UK
7
+ 3 Institute of Computing Technology, Chinese Academy of Sciences, China
8
+ 4 University of Michigan Ann Arbor, USA
9
+ 5 University of Colorado Boulder, USA
10
+ 6 Sichuan University, China
11
+ 7 McGill University, Canada
12
+ {yizhi.li, c.lin}@sheffield.ac.uk2, [email protected], [email protected]
13
+ Abstract
14
+ As natural language processing (NLP) for gen-
15
+ der bias becomes a significant interdisciplinary
16
+ topic, the prevalent data-driven techniques,
17
+ such as large-scale language models, suffer
18
+ from data inadequacy and biased corpus, espe-
19
+ cially for languages with insufficient resources,
20
+ such as Chinese.
21
+ To this end, we propose
22
+ a Chinese cOrpus foR Gender bIas Probing
23
+ and Mitigation (CORGI-PM1), which con-
24
+ tains 32.9k sentences with high-quality labels
25
+ derived by following an annotation scheme
26
+ specifically developed for gender bias in the
27
+ Chinese context. Moreover, we address three
28
+ challenges for automatic textual gender bias
29
+ mitigation, which requires the models to de-
30
+ tect, classify, and mitigate textual gender bias.
31
+ We also conduct experiments with state-of-the-
32
+ art language models to provide baselines. To
33
+ our best knowledge, CORGI-PM is the first
34
+ sentence-level Chinese corpus for gender bias
35
+ probing and mitigation.
36
+ 1
37
+ Introduction
38
+ Increasing recognition in consensus is that iden-
39
+ tifying and preventing toxic gender attitudes and
40
+ stereotypes is essential for society (Blodgett et al.,
41
+ 2020). Since gender-biased information could be
42
+ presented and widely propagated in textual format,
43
+ it is essential to develop automatic methods for
44
+ detecting and mitigating textual gender bias.
45
+ Natural language processing (NLP) has been
46
+ widely used in text-related applications, which have
47
+ a significant influence on gender bias topics (Costa-
48
+ jussà, 2019). On the one hand, large-scale language
49
+ models (LMs), as a key technique of modern NLP,
50
+ are proven to learn the subjective gender bias in
51
+ the training corpus or even amplify it (Zhao et al.,
52
+ 2017) On the other hand, it becomes increasingly
53
+ ∗ The two authors contributed equally to this work.
54
+ † Corresponding authors.
55
+ 1Our code is available at GitHub
56
+ promising to apply cutting-edge NLP techniques
57
+ for probing and mitigating gender bias.
58
+ Language
59
+ Models
60
+ Gender-related
61
+ Vocabulary
62
+ Polarity
63
+ Calculation
64
+ Word
65
+ Matching
66
+ Sentence-level
67
+ Reranking
68
+ Potentially
69
+ Biased Corpus
70
+ Corpus with
71
+ Biased Vocabulary
72
+ Raw Corpus
73
+ Figure 1: Pipeline of Retrieving and Filtering Potentially Bi-
74
+ ased Sentences from Raw Corpus for Human Annotation.
75
+ Building a high-quality text corpus has been one
76
+ of the key tangents in improving NLP applications
77
+ for debiasing gender stereotypes in texts (Sun et al.,
78
+ 2019). Some researchers introduce automatic an-
79
+ notation techniques, such as gender-swapped based
80
+ methods, to create corpora for gender bias mitiga-
81
+ tion (Lu et al., 2020; Zhao et al., 2018; Rudinger
82
+ et al., 2018). While it is attractive to build a large
83
+ corpus without heavy labors, automatic gender-
84
+ swapped based methods highly depend on the qual-
85
+ ity of base language models and are prone to cre-
86
+ ating nonsensical sentences (Sun et al., 2019). To
87
+ address this issue, some works devote effort to de-
88
+ veloping human-annotated corpora for gender bias
89
+ mitigation. However, these corpora either mainly
90
+ focus on word- or grammar-level bias (Webster
91
+ et al., 2018; Zhu and Liu, 2020; Sahai and Sharma,
92
+ 2021; Zhou et al., 2019), or only concern about
93
+ sexism-related topics (Jiang et al., 2022; Chiril
94
+ et al., 2021, 2020; Parikh et al., 2019).
95
+ Moreover, existing works on gender bias exclu-
96
+ arXiv:2301.00395v1 [cs.CL] 1 Jan 2023
97
+
98
+ sively focus on English (Costa-jussà, 2019), where
99
+ few datasets exist for other influential languages
100
+ such as Chinese. (N.B. details of generated gen-
101
+ der bias corpus with nonsensical Chinese sentences
102
+ can be found in Appendix D). We aim to tackle the
103
+ aforementioned issues by providing a high-quality
104
+ Chinese human-annotated corpus for contextual-
105
+ level gender bias probing and mitigation.
106
+ To this end, we propose the Chinese cOrpus foR
107
+ Gender bIas Probing and Mitigation (CORGI-PM)
108
+ dataset, which consists of 32.9k human-annotated
109
+ sentences, including both gender-biased and non-
110
+ biased samples.
111
+ For the initial data collection,
112
+ we propose an automatic method that builds a po-
113
+ tentially gender-biased sentence set from existing
114
+ large-scale Chinese corpora. Inspired by the metric
115
+ leveraging language models for gender bias score
116
+ calculation proposed in Bolukbasi et al. (2016);
117
+ Jiao and Luo (2021), the samples containing words
118
+ of high gender bias scores are recalled, and then
119
+ reranked and filtered according to their sentence-
120
+ level gender-biased probability, as illustrated in
121
+ Fig. 1. To ensure the quality of our corpus, the
122
+ annotation scheme is carefully designed, and an-
123
+ notators with qualified educational backgrounds
124
+ are selected to further label and paraphrase the re-
125
+ trieved sentences.
126
+ Additionally, we address three challenges based
127
+ on CORGI-PM, i.e., gender bias detection, classifi-
128
+ cation, and mitigation, which provide clear defini-
129
+ tions and evaluation protocols for NLP tasks in gen-
130
+ der bias probing and mitigation. In order to provide
131
+ referential baselines and benchmarks for our pro-
132
+ posed challenges, we conduct random data splitting
133
+ with balanced labels and implement experiments
134
+ on cutting-edge language models in zero-shot, in-
135
+ context learning, and fine-tuning paradigms. We
136
+ discuss the experimental settings and provide result
137
+ analysis in §3. The implementation details can be
138
+ referred to in Appendix C.
139
+ In summary, we provide a well-annotated Chi-
140
+ nese corpus for gender bias probing and mitiga-
141
+ tion, along with clearly defined corresponding
142
+ challenges. With a properly designed annotation
143
+ scheme, CORGI-PM provides a corpus of high
144
+ quality that assists models in detecting gender bias
145
+ in texts. More importantly, other than the 22.5k
146
+ human-annotated non-biased samples, all the 5.2k
147
+ biased sentences in our corpus are further labeled
148
+ with gender bias subclasses and companies with
149
+ parallel bias-free versions provided by the annota-
150
+ Sample
151
+ Quantity
152
+ Category
153
+ Train
154
+ Valid
155
+ Test
156
+ Biased
157
+ AC
158
+ 1.90k
159
+ 235
160
+ 237
161
+ DI
162
+ 2.70k
163
+ 334
164
+ 337
165
+ ANB
166
+ 2.47k
167
+ 306
168
+ 309
169
+ Non-biased
170
+ 21.4k
171
+ 516
172
+ 526
173
+ Overall
174
+ 30.1k
175
+ 1391
176
+ 1409
177
+ Table 1: Overall Statistics of the CORGI-PM Dataset. The
178
+ notations, AC, DI, and ANB represent specific bias labels
179
+ described in § 2.2.
180
+ tors. Our codes and dataset will be released for the
181
+ benefit of the community.
182
+ 2
183
+ Data Collection
184
+ 2.1
185
+ Sample Filtering
186
+ We propose an automatic processing method to
187
+ recall, rerank, and filter annotation candidates from
188
+ raw corpora using a two-stage filtering from word-
189
+ level to sentence-level, as illustrated in Fig. 1. The
190
+ Chinese sentence samples are mainly screened out
191
+ from the SlguSet (Zhao et al., 2021) and the CCL
192
+ corpus (Weidong et al., 2019).
193
+ To recall gender-biased words or retrieve candi-
194
+ date sentences with gender bias scores, we com-
195
+ pare the target word/sentence representations with
196
+ the seed direction, which can be calculated by the
197
+ subtraction between the word embeddings of she
198
+ and he
199
+ (Bolukbasi et al., 2016; Jiao and Luo,
200
+ 2021). We leverage different Chinese LMs includ-
201
+ ing ERNIE (Zhang et al., 2019), CBert (Cui et al.,
202
+ 2020), and Chinese word vectors (Qiu et al., 2018)
203
+ to acquire the word-level and sentence-level rep-
204
+ resentations. For word-level filtering, we use the
205
+ mentioned metric to build a vocabulary of high
206
+ bias scores and recall sentences containing such
207
+ words from the raw corpora with exact matches.
208
+ We compute gender bias scores of the crawled sen-
209
+ tences and group them by the gender bias keywords
210
+ acquired in the previous stage for sentence-level fil-
211
+ tering. The final sentences for annotation are then
212
+ selected according to a specific global threshold
213
+ gender bias score and an in-group threshold rank.
214
+ The word-level filtering process presented as word
215
+ clouds can be found in Appendix B.1.
216
+ 2.2
217
+ Annotation Scheme
218
+ The annotation scheme is designed for gender bias
219
+ probing and mitigation. For gender bias probing,
220
+ the annotators are required to provide the follow-
221
+ ing information given a sentence: whether gender
222
+ bias exists; if so, how the bias is established. For
223
+ gender bias mitigation, the corrected non-biased
224
+ version of the biased sentences is also required. We
225
+
226
+ Linguistic
227
+ Non-biased
228
+ Biased
229
+ Corrected Biased
230
+ Info.
231
+ Train
232
+ Valid
233
+ Test
234
+ Train
235
+ Valid
236
+ Test
237
+ Train
238
+ Valid
239
+ Test
240
+ Word
241
+ 724k
242
+ 18.9k
243
+ 17.7k
244
+ 228k
245
+ 24.8k
246
+ 28.3k
247
+ 265k
248
+ 27.1k
249
+ 30.0k
250
+ Dictionary
251
+ 574k
252
+ 14.4k
253
+ 14.1k
254
+ 167k
255
+ 18.4k
256
+ 20.4k
257
+ 191k
258
+ 19.9k
259
+ 21.5k
260
+ Character
261
+ 1,156k
262
+ 30.1k
263
+ 28.1k
264
+ 358k
265
+ 39.2k
266
+ 44.4k
267
+ 417k
268
+ 42.8k
269
+ 46.9k
270
+ Sent. Length
271
+ 53.952
272
+ 58.397
273
+ 53.473
274
+ 85.837
275
+ 76.087
276
+ 85.214
277
+ 99.839
278
+ 82.853
279
+ 89.939
280
+ Table 2: Linguistic Characteristics of the Corpus. Word, Dictionary, and Character separately denote the total Chinese word
281
+ number, total unique Chinese word number, and total character number of the specific categories. The sentence lengths are
282
+ defined as the number of containing characters.
283
+ further describe the annotation scheme details in
284
+ the following paragraphs.
285
+ Existence and Categorization.
286
+ The annotators are required to annotate whether
287
+ the sentence is gender-biased (B) or non-biased (N)
288
+ in contextual-level or word-level, and further clar-
289
+ ify how the bias is established. Given that our raw
290
+ data is collected using gender-related keywords or
291
+ from gender-related corpus, the samples annotated
292
+ without gender bias are useful human-annotated
293
+ negative samples for detecting gender bias. To addi-
294
+ tionally provide information about gender bias cate-
295
+ gorization, we classify gender bias types into three
296
+ subtypes : (1) Gender Stereotyped activity and
297
+ career choices (AC); (2) Gender Stereotyped de-
298
+ scriptions and inductions (DI); and (3) Expressed
299
+ gender-stereotyped attitudes, norms and beliefs
300
+ (ANB). The classification standard is inspired by
301
+ (King et al., 2021) and further summed up into the
302
+ mentioned subtypes.
303
+ Bias Mitigation. Annotators are also required to
304
+ mitigate the gender bias of selected sentences while
305
+ keeping the original semantic information. We
306
+ also ask our annotators to diversify the expres-
307
+ sions if applicable. The major revision patterns
308
+ can be summarized as follows: (1). Replace the
309
+ gender-specific pronouns with neutral pronouns.
310
+ (2). Replace the gender-specific adjectives with
311
+ neutral descriptions with similar semantics defini-
312
+ tions. (3). Add additional comments to neutralize
313
+ the sentences which cannot be directly mitigated.
314
+ 2.3
315
+ Corpus Analysis
316
+ In this section, we report the linguistic statistics
317
+ of CORGI-PM as Tab. 1. We design a balanced
318
+ split to create the valid and test set considering the
319
+ negative-positive ratio and bias subclass proportion
320
+ in the global distribution. As revealed in Tab. 22,
321
+ we observe two major differences compared the de-
322
+ biased samples with the original ones: longer and
323
+ more diverse expressions (N.B. sentence length and
324
+ vocabulary size of Tab. 2). We hypothesize that it
325
+ 2We use the Jieba to parse.
326
+ is due to human annotators’ intention to keep the
327
+ semantic information unchanged and the sentence
328
+ coherent while mitigating gender bias. They may
329
+ use more conjunctions and longer descriptions com-
330
+ pared to some gender-biased inherent expressions.
331
+ More details for quality managing and control can
332
+ be referred to Appendix B.1 and B.2.
333
+ 3
334
+ Gender Bias Mitigation Challenges
335
+ To provide a clear definition for automatic textual
336
+ gender bias probing and mitigation tasks, we pro-
337
+ pose corresponding challenges and standardize the
338
+ evaluation protocols. We address two tasks, detec-
339
+ tion, and classification, for gender bias probing and
340
+ formalize the gender mitigation challenge as a text
341
+ mitigation task.
342
+ 3.1
343
+ Challenges of Detection and
344
+ Classification
345
+ We regard both the gender bias detection and clas-
346
+ sification challenges as supervised classification
347
+ tasks and evaluate them with metrics of consensus.
348
+ Definition. The gender bias detection challenge
349
+ can be regarded as a binary classification task,
350
+ where the model is required to predict the prob-
351
+ ability that a given sentence contains gender bias.
352
+ As described in § 2.2, biased samples are further
353
+ categorized into one or more kinds. Therefore, we
354
+ can address the gender classification challenge as
355
+ a multi-label classification task. The precision, re-
356
+ call, and F1-score are selected as the main metrics
357
+ in these two challenges. Class-wise metrics and
358
+ macro average summarized evaluation are required
359
+ through both valid and test sets to show the perfor-
360
+ mance of language models.
361
+ Baselines. We finetune Chinese language mod-
362
+ els from three representative different pretrained
363
+ paradigms, i.e., Chinese BERT, Electra, and XL-
364
+ Net Cui et al. (2020), for both the detection and
365
+ classification tasks by adding an additional dense
366
+ prediction layer. 3 We also provide GPT-3 (Brown
367
+ et al., 2020) curie’s few-shot performance for both
368
+ 3Pretrained models can be found at theHFL Anthology.
369
+
370
+ Model
371
+ Metrics
372
+ Classification (Val.)
373
+ Classification (Test)
374
+ Detection (Val.)
375
+ Detection (Test.)
376
+ AC
377
+ DI
378
+ ANB
379
+ Avg.
380
+ AC
381
+ DI
382
+ ANB
383
+ Avg.
384
+ N
385
+ B
386
+ Avg.
387
+ N
388
+ B
389
+ Avg.
390
+ BERT
391
+ Precision
392
+ .609
393
+ .729
394
+ .533
395
+ .624
396
+ .556
397
+ .615
398
+ .521
399
+ .564
400
+ .699
401
+ .950
402
+ .824
403
+ .742
404
+ .980
405
+ .861
406
+ Recall
407
+ .594
408
+ .665
409
+ .543
410
+ .601
411
+ .493
412
+ .652
413
+ .585
414
+ .577
415
+ .971
416
+ .591
417
+ .781
418
+ .985
419
+ .662
420
+ .823
421
+ F1-Score
422
+ .602
423
+ .695
424
+ .538
425
+ .612
426
+ .522
427
+ .633
428
+ .551
429
+ .567
430
+ .813
431
+ .729
432
+ .771
433
+ .846
434
+ .790
435
+ .818
436
+ Electra
437
+ Precision
438
+ .587
439
+ .727
440
+ .544
441
+ .619
442
+ .556
443
+ .630
444
+ .516
445
+ .568
446
+ .691
447
+ .936
448
+ .814
449
+ .745
450
+ .974
451
+ .860
452
+ Recall
453
+ .758
454
+ .687
455
+ .386
456
+ .610
457
+ .680
458
+ .685
459
+ .373
460
+ .579
461
+ .961
462
+ .570
463
+ .766
464
+ .983
465
+ .656
466
+ .820
467
+ F1-Score
468
+ .661
469
+ .706
470
+ .451
471
+ .606
472
+ .612
473
+ .656
474
+ .433
475
+ .567
476
+ .804
477
+ .708
478
+ .756
479
+ .848
480
+ .784
481
+ .816
482
+ XLNet
483
+ Precision
484
+ .587
485
+ .696
486
+ .523
487
+ .602
488
+ .544
489
+ .589
490
+ .527
491
+ .553
492
+ .713
493
+ .928
494
+ .820
495
+ .772
496
+ .959
497
+ .865
498
+ Recall
499
+ .622
500
+ .643
501
+ .495
502
+ .587
503
+ .545
504
+ .614
505
+ .514
506
+ .558
507
+ .953
508
+ .620
509
+ .787
510
+ .968
511
+ .722
512
+ .845
513
+ F1-Score
514
+ .604
515
+ .669
516
+ .509
517
+ .594
518
+ .544
519
+ .601
520
+ .520
521
+ .555
522
+ .816
523
+ .743
524
+ .780
525
+ .859
526
+ .824
527
+ .841
528
+ Curie
529
+ Precision
530
+ .695
531
+ .907
532
+ .010
533
+ .537
534
+ .622
535
+ .887
536
+ .009
537
+ .506
538
+ .763
539
+ .665
540
+ .714
541
+ .635
542
+ .825
543
+ .730
544
+ Recall
545
+ .395
546
+ .802
547
+ .375
548
+ .524
549
+ .395
550
+ .804
551
+ .010
552
+ .403
553
+ .576
554
+ .825
555
+ .700
556
+ .975
557
+ .584
558
+ .780
559
+ F1-Score
560
+ .504
561
+ .851
562
+ .019
563
+ .458
564
+ .508
565
+ .852
566
+ .019
567
+ .460
568
+ .656
569
+ .736
570
+ .696
571
+ .769
572
+ .684
573
+ .727
574
+ Table 3: Baseline Results for Gender Bias Detection and Classification Tasks. The overall metric refers to Marco average. The
575
+ model names and abbreviations refer to § 3.1. Categorical definitions refer to § 2.2.
576
+ aa
577
+ Metrics
578
+ Models
579
+ BLEU
580
+ METEOR
581
+ ROUGE-L
582
+ Human Evaluation
583
+ Recall
584
+ Precision
585
+ F1
586
+ Coherence
587
+ Gender Bias
588
+ *Davinci
589
+ .776
590
+ .879
591
+ .203
592
+ .211
593
+ .205
594
+ 5.25
595
+ 0.96
596
+ Ada
597
+ .288
598
+ .429
599
+ .407
600
+ .180
601
+ .250
602
+ 5.98
603
+ 1.13
604
+ Babbage
605
+ .359
606
+ .504
607
+ .716
608
+ .310
609
+ .432
610
+ 6.32
611
+ 0.69
612
+ Curie
613
+ .364
614
+ .506
615
+ .692
616
+ .316
617
+ .434
618
+ 6.21
619
+ 1.20
620
+ Table 4: Baseline Results for Gender Bias Correction task. Metrics details can be found in Appendix C. * suggests using the
621
+ model in zero-shot paradigm and the others refers to fine-tune.
622
+ the detection and classification tasks. Baseline re-
623
+ sults of detection and classification show that the
624
+ classification task is challenging, and there is room
625
+ for performance improvement in detecting gender
626
+ bias in CORGI-PM, as revealed in Tab. 3.
627
+ 3.2
628
+ Challenge of Mitigation
629
+ Definition. The gender bias mitigation challenge
630
+ can be regarded as a natural language generation
631
+ task, where the model is asked to generate the cor-
632
+ rected version of biased sentences with the human-
633
+ annotated ones as references.
634
+ Baselines. We test the GPT-3 (Brown et al., 2020)
635
+ on CORGI-PM in fine-tune experiment setting
636
+ with three different parameter scales, which are
637
+ Ada(350M), Babbage(1.3B), and Curie(6.7B), and
638
+ Davinci(175B) in zero-shot experiment setting. We
639
+ only provide zero-shot results for Davinci because
640
+ it is the only released GPT-3 editing model. More
641
+ implementation and evaluation details are intro-
642
+ duced in Appendix C.
643
+ Discussion. We provide both human evaluation
644
+ and automated metrics for evaluation. Tab. 4 re-
645
+ veals that LMs can learn the annotation pattern of
646
+ mitigating gender bias, and the zero-shot editing
647
+ model shows competitive performance. The obser-
648
+ vation that fine-tuned Babbage outperforms much
649
+ larger zero-shot Davinci in the human evaluation,
650
+ and ROUGE-L reveals that CORGI-PM has the
651
+ potential to be used as strong supervision of the
652
+ gender bias mitigation task. We notice that Davinci
653
+ tends to apply more conservative edits compared to
654
+ fine-tuned models. As a result, the sentences edited
655
+ by Davinci keep most of the original sentences and
656
+ always only change pronouns and adjectives from
657
+ the original sentences, which benefits precision
658
+ focusing automatic metrics like BLEU (Papineni
659
+ et al., 2002), and METEOR (Agarwal and Lavie,
660
+ 2007). The performance difference between human
661
+ evaluation and automatic metrics reveals the writ-
662
+ ing style difference between human and language
663
+ models.
664
+ 4
665
+ Conclusion
666
+ We propose CORGI-PM, the first Chinese human-
667
+ annotated corpus for both gender bias probing and
668
+ mitigation. We also address definitions and evalua-
669
+ tion metrics for three challenges based on CORGI-
670
+ PM and test the performances of state-of-the-art
671
+ language models. Our proposed challenges can
672
+ serve as benchmarks for measuring the ability of
673
+ language models to detect, classify, and mitigate
674
+ textual gender bias. Experiments show that our sen-
675
+ tences with fine-grained subclass labels can assist
676
+ the language models in gender bias probing, whilst
677
+ our parallel human-written debiased data can serve
678
+ as strong supervision of the generative language
679
+ models. In summary, we imply future work utiliz-
680
+ ing CORGI-PM would be benefited the topic of
681
+ NLP for gender bias probing and mitigation.
682
+ Limitations
683
+ There are several major limitations in this research
684
+ work. Due to the high requirement of annotators
685
+
686
+ for annotating gender-biased sentences and correct-
687
+ ing such sentences, we only choose annotators with
688
+ higher education, which may lead to potential cog-
689
+ nitive bias. In addition, we only conduct limited
690
+ implementations and experiments of testing widely-
691
+ used Chinese language models’ performance in our
692
+ new challenges. More language models and tech-
693
+ niques can be further explored in our challenges.
694
+ Ethics Statement
695
+ We carefully consider the ethical implications dur-
696
+ ing the collection process. The collection of our
697
+ corpus CORGI-PM sentences only relies on public
698
+ available corpora for research purposes. We have
699
+ acknowledged the potential usage of our dataset as
700
+ well as related privacy issues to the annotators and
701
+ received confirmations before the annotation was
702
+ initiated.
703
+ References
704
+ Abhaya Agarwal and Alon Lavie. 2007. Meteor: An
705
+ automatic metric for mt evaluation with high levels
706
+ of correlation with human judgments. Proceedings
707
+ of WMT-08.
708
+ Su Lin Blodgett, Solon Barocas, Hal Daumé III, and
709
+ Hanna Wallach. 2020.
710
+ Language (technology) is
711
+ power: A critical survey of “bias” in nlp. In ACL.
712
+ Tolga Bolukbasi, Kai-Wei Chang, James Y. Zou,
713
+ Venkatesh Saligrama, and Adam Tauman Kalai.
714
+ 2016. Man is to computer programmer as woman
715
+ is to homemaker? debiasing word embeddings. In
716
+ NIPS.
717
+ Tom Brown, Benjamin Mann, Nick Ryder, Melanie
718
+ Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind
719
+ Neelakantan, Pranav Shyam, Girish Sastry, Amanda
720
+ Askell, et al. 2020. Language models are few-shot
721
+ learners. Advances in neural information processing
722
+ systems, 33:1877–1901.
723
+ Patricia Chiril,
724
+ Farah Benamara,
725
+ and Véronique
726
+ Moriceau. 2021. “be nice to your wife! the restau-
727
+ rants are closed”: Can gender stereotype detection
728
+ improve sexism classification?
729
+ In Findings of the
730
+ Association for Computational Linguistics: EMNLP
731
+ 2021, pages 2833–2844.
732
+ Patricia Chiril, Véronique Moriceau, Farah Benamara,
733
+ Alda Mari, Gloria Origgi, and Marlène Coulomb-
734
+ Gully. 2020. An annotated corpus for sexism detec-
735
+ tion in french tweets. In Proceedings of the 12th lan-
736
+ guage resources and evaluation conference, pages
737
+ 1397–1403.
738
+ Marta R Costa-jussà. 2019. An analysis of gender bias
739
+ studies in natural language processing. Nature Ma-
740
+ chine Intelligence, 1(11):495–496.
741
+ Yiming Cui, Wanxiang Che, Ting Liu, Bing Qin, Shi-
742
+ jin Wang, and Guoping Hu. 2020. Revisiting pre-
743
+ trained models for chinese natural language process-
744
+ ing. arXiv preprint arXiv:2004.13922.
745
+ Aiqi Jiang, Xiaohan Yang, Yang Liu, and Arkaitz Zu-
746
+ biaga. 2022. Swsr: A chinese dataset and lexicon
747
+ for online sexism detection. Online Social Networks
748
+ and Media, 27:100182.
749
+ Meichun Jiao and Ziyang Luo. 2021. Gender bias hid-
750
+ den behind chinese word embeddings: The case of
751
+ chinese adjectives. In Proceedings of the 3rd Work-
752
+ shop on Gender Bias in Natural Language Process-
753
+ ing, pages 8–15.
754
+ Tania L King, Anna J Scovelle, Anneke Meehl, Al-
755
+ lison J Milner, and Naomi Priest. 2021.
756
+ Gender
757
+ stereotypes and biases in early childhood: A sys-
758
+ tematic review. Australasian Journal of Early Child-
759
+ hood, 46(2):112–125.
760
+ Chin-Yew Lin. 2004. Rouge: A package for automatic
761
+ evaluation of summaries.
762
+ In Text summarization
763
+ branches out, pages 74–81.
764
+ Kaiji Lu, Piotr Mardziel, Fangjing Wu, Preetam Aman-
765
+ charla, and Anupam Datta. 2020.
766
+ Gender bias in
767
+ neural natural language processing. In Logic, Lan-
768
+ guage, and Security, pages 189–202. Springer.
769
+ Kishore Papineni, Salim Roukos, Todd Ward, and Wei-
770
+ Jing Zhu. 2002. Bleu: a method for automatic eval-
771
+ uation of machine translation. In Proceedings of the
772
+ 40th annual meeting of the Association for Compu-
773
+ tational Linguistics, pages 311–318.
774
+ Pulkit Parikh, Harika Abburi, Pinkesh Badjatiya, Rad-
775
+ hika Krishnan, Niyati Chhaya, Manish Gupta, and
776
+ Vasudeva Varma. 2019. Multi-label categorization
777
+ of accounts of sexism using a neural framework.
778
+ arXiv preprint arXiv:1910.04602.
779
+ Yuanyuan Qiu, Hongzheng Li, Shen Li, Yingdi Jiang,
780
+ Renfen Hu, and Lijiao Yang. 2018. Revisiting cor-
781
+ relations between intrinsic and extrinsic evaluations
782
+ of word embeddings.
783
+ In Chinese Computational
784
+ Linguistics and Natural Language Processing Based
785
+ on Naturally Annotated Big Data, pages 209–221.
786
+ Springer.
787
+ Rachel Rudinger, Jason Naradowsky, Brian Leonard,
788
+ and
789
+ Benjamin
790
+ Van
791
+ Durme.
792
+ 2018.
793
+ Gender
794
+ bias in coreference resolution.
795
+ arXiv preprint
796
+ arXiv:1804.09301.
797
+ Saumya Sahai and Dravyansh Sharma. 2021. Predict-
798
+ ing and explaining french grammatical gender. In
799
+ Proceedings of the Third Workshop on Computa-
800
+ tional Typology and Multilingual NLP, pages 90–96.
801
+ Aarohi Srivastava, Abhinav Rastogi, Abhishek Rao,
802
+ Abu Awal Md Shoeb, Abubakar Abid, Adam Fisch,
803
+ Adam R Brown, Adam Santoro, Aditya Gupta,
804
+ Adrià Garriga-Alonso, et al. 2022.
805
+ Beyond the
806
+ imitation game: Quantifying and extrapolating the
807
+
808
+ capabilities of language models.
809
+ arXiv preprint
810
+ arXiv:2206.04615.
811
+ Tony Sun, Andrew Gaut, Shirlyn Tang, Yuxin Huang,
812
+ Mai ElSherief, Jieyu Zhao, Diba Mirza, Eliza-
813
+ beth Belding, Kai-Wei Chang, and William Yang
814
+ Wang. 2019. Mitigating gender bias in natural lan-
815
+ guage processing: Literature review. arXiv preprint
816
+ arXiv:1906.08976.
817
+ Zeerak Waseem. 2016. Are you a racist or am I seeing
818
+ things? annotator influence on hate speech detection
819
+ on Twitter. In Proceedings of the First Workshop on
820
+ NLP and Computational Social Science, pages 138–
821
+ 142, Austin, Texas. Association for Computational
822
+ Linguistics.
823
+ Kellie Webster, Marta Recasens, Vera Axelrod, and Ja-
824
+ son Baldridge. 2018.
825
+ Mind the gap: A balanced
826
+ corpus of gendered ambiguous pronouns. Transac-
827
+ tions of the Association for Computational Linguis-
828
+ tics, 6:605–617.
829
+ Zhan Weidong, Guo Rui, Chang Baobao, Chen Yirong,
830
+ and Long Chen. 2019. Development of Peking Uni-
831
+ versity CCL Corpus. In Chinese Corpus Linguistic
832
+ Journal.
833
+ Huimin Xu, Zhang Zhang, Lingfei Wu, and Cheng-Jun
834
+ Wang. 2019. The cinderella complex: Word embed-
835
+ dings reveal gender stereotypes in movies and books.
836
+ PloS one, 14(11):e0225385.
837
+ Zhengyan Zhang, Xu Han, Zhiyuan Liu, Xin Jiang,
838
+ Maosong Sun, and Qun Liu. 2019.
839
+ Ernie:
840
+ En-
841
+ hanced language representation with informative en-
842
+ tities. arXiv preprint arXiv:1905.07129.
843
+ Jieyu Zhao, Tianlu Wang, Mark Yatskar, Vicente
844
+ Ordonez, and Kai-Wei Chang. 2017.
845
+ Men also
846
+ like shopping:
847
+ Reducing gender bias amplifica-
848
+ tion using corpus-level constraints. arXiv preprint
849
+ arXiv:1707.09457.
850
+ Jieyu Zhao, Yichao Zhou, Zeyu Li, Wei Wang, and Kai-
851
+ Wei Chang. 2018.
852
+ Learning gender-neutral word
853
+ embeddings. arXiv preprint arXiv:1809.01496.
854
+ Jishun Zhao, Bingjie Du, Shucheng Zhu, and Pengyuan
855
+ Liu. 2021. Construction of chinese sentence-level
856
+ gender-unbiased data set and evaluation of gender
857
+ bias in pre-training language.
858
+ In Proceedings of
859
+ the 20th Chinese National Conference on Computa-
860
+ tional Linguistics, pages 564–575.
861
+ Pei
862
+ Zhou,
863
+ Weijia
864
+ Shi,
865
+ Jieyu
866
+ Zhao,
867
+ Kuan-Hao
868
+ Huang, Muhao Chen, Ryan Cotterell, and Kai-
869
+ Wei Chang. 2019.
870
+ Examining gender bias in lan-
871
+ guages with grammatical gender.
872
+ arXiv preprint
873
+ arXiv:1909.02224.
874
+ Shucheng Zhu and Pengyuan Liu. 2020. Great males
875
+ and stubborn females: A diachronic study of corpus-
876
+ based gendered skewness in chinese adjectives. In
877
+ Proceedings of the 19th Chinese National Confer-
878
+ ence on Computational Linguistics, pages 31–42.
879
+
880
+ ERNIE
881
+ BERT
882
+ XLNet
883
+ ELECTRA
884
+ ERNIE
885
+ BERT
886
+ XLNet
887
+ ELECTRA
888
+ 1
889
+ -0.015
890
+ 0.058
891
+ -0.0075
892
+ -0.015
893
+ 1
894
+ 0.021
895
+ -0.083
896
+ 0.058
897
+ 0.021
898
+ 1
899
+ 0.14
900
+ -0.0075
901
+ -0.083
902
+ 0.14
903
+ 1
904
+ 0.0
905
+ 0.2
906
+ 0.4
907
+ 0.6
908
+ 0.8
909
+ 1.0
910
+ Figure 2: Word-level Gender Bias Comparison of Career
911
+ Words.
912
+ A
913
+ Gender Bias Analysis of Chinese
914
+ Language Models
915
+ A.1
916
+ Evaluation Method and Data Sets
917
+ We conduct experiments to explore gender bias con-
918
+ tained in widely-used Chinese language models for
919
+ research and industrial use. We employ the method
920
+ Bolukbasi et al. (2016); Jiao and Luo (2021) pro-
921
+ posed to assess gender bias. The gender bias score
922
+ for a word is calculated by ⃗w · ( ⃗
923
+ she − ⃗he)based on
924
+ its word vector. A positive value means the word
925
+ is more relevant to females, while a negative value
926
+ means the word is more relevant to males. The
927
+ higher the absolute value of the gender bias score,
928
+ the more biased the word indicates.
929
+ Srivastava et al. propose a big benchmark con-
930
+ taining a dataset specifying the existing Chinese ca-
931
+ reer words. Zhu and Liu propose AGSS, a manual-
932
+ created Chinese word-level adjective list containing
933
+ gender bias. To measure gender bias contained in
934
+ the language models, we first calculate gender bias
935
+ scores of words in the word list provided (Srivas-
936
+ tava et al., 2022; Zhu and Liu, 2020) according to
937
+ the projection method Bolukbasi et al. (2016); Jiao
938
+ and Luo (2021). We compare the career and adjec-
939
+ tive word gender bias score vectors to get the ob-
940
+ servations of LMs’ influence on word-level learned
941
+ gender bias. To make the observations more clear,
942
+ we further apply the sign function to the career and
943
+ adjective word gender bias score vectors. The sim-
944
+ ilarity function used for the heatmaps is Pearson
945
+ similarity.
946
+ We conduct described comparison of adjectives
947
+ between AGSS as a golden standard (Zhu and Liu,
948
+ 2020), Ernie (Zhang et al., 2019), Chinese Word
949
+ Vectors trained by mixed corpus (Qiu et al., 2018),
950
+ AGSS
951
+ ERNIE
952
+ BERT
953
+ CWV
954
+ XLNet
955
+ ELECTRA
956
+ AGSS
957
+ ERNIE
958
+ BERT
959
+ CWV
960
+ XLNet
961
+ ELECTRA
962
+ 1
963
+ 0.22
964
+ -0.023
965
+ 0.15
966
+ 0.021 0.044
967
+ 0.22
968
+ 1
969
+ -0.037 0.066 0.012
970
+ 0.1
971
+ -0.023 -0.037
972
+ 1
973
+ 0.074 0.032 -0.0097
974
+ 0.15
975
+ 0.066 0.074
976
+ 1
977
+ -0.065 0.017
978
+ 0.021 0.012 0.032 -0.065
979
+ 1
980
+ 0.0093
981
+ 0.044
982
+ 0.1
983
+ -0.0097 0.017 0.0093
984
+ 1
985
+ 0.0
986
+ 0.2
987
+ 0.4
988
+ 0.6
989
+ 0.8
990
+ 1.0
991
+ Figure 3: Word-level Gender Bias Comparison of Adjectives.
992
+ CWV denotes the Chinese Word Vectors trained using mixed-
993
+ large corpus proposed by Qiu et al..
994
+ AGSS
995
+ Mixed-large
996
+ PDN
997
+ Zhihu_QA
998
+ Weibo
999
+ Literature
1000
+ AGSS
1001
+ Mixed-large
1002
+ PDN
1003
+ Zhihu_QA
1004
+ Weibo
1005
+ Literature
1006
+ 1
1007
+ 0.21
1008
+ 0.15
1009
+ -0.093 0.086
1010
+ 0.21
1011
+ 0.21
1012
+ 1
1013
+ 0.071
1014
+ -0.25 -0.043 -0.021
1015
+ 0.15
1016
+ 0.071
1017
+ 1
1018
+ -0.041 -0.025 0.046
1019
+ -0.093 -0.25 -0.041
1020
+ 1
1021
+ 0.12
1022
+ -0.037
1023
+ 0.086 -0.043 -0.025
1024
+ 0.12
1025
+ 1
1026
+ -0.014
1027
+ 0.21
1028
+ -0.021 0.046 -0.037 -0.014
1029
+ 1
1030
+ 0.2
1031
+ 0.0
1032
+ 0.2
1033
+ 0.4
1034
+ 0.6
1035
+ 0.8
1036
+ 1.0
1037
+ Figure 4: Word-level Gender Bias Comparison of Adjectives
1038
+ of Language Models Pre-trained by Different Corpus. PDN
1039
+ denotes the People’s Daily News Corpus.
1040
+ and Chinese-XLNet, Chinese-Bert, and Chinese-
1041
+ Electra proposed tecui-etal-2020-revisiting to pro-
1042
+ duce Fig. 3. We conduct described comparison of
1043
+ career words between Ernie (Zhang et al., 2019),
1044
+ and Chinese-XLNet, Chinese-Bert, and Chinese-
1045
+ Electra proposed tecui-etal-2020-revisiting to pro-
1046
+ duce Fig. 2. The described experiments on career
1047
+ words is not conducted with the Chinese Word Vec-
1048
+ tors trained by mixed corpus, because an observing
1049
+ number of career words are missing in its dictio-
1050
+ nary.
1051
+ We don’t provide a golden standard vector (Sri-
1052
+ vastava et al., 2022) since they didn’t provide a
1053
+ manual gender bias analysis about the career words.
1054
+ We also conduct described comparison on adjec-
1055
+ tives in Chinese Word Vectors pre-trained by dif-
1056
+ ferent corpus, including Mixed-large corpus, Peo-
1057
+ ple’s Daily News, Zhihu QA dataset, Weibo, and
1058
+ Chinese literature dataset to produce Fig. 4 and an-
1059
+ alyze the learned gender bias difference caused by
1060
+
1061
+ (a) Ch-Ernie-Man-Adj
1062
+ (b) Ch-Ernie-Woman-Adj
1063
+ (c) Ch-Ernie-Man-Career
1064
+ (d) Ch-Ernie-Woman-Career
1065
+ (e) En-Ernie-Man-Adj
1066
+ (f) En-Ernie-Woman-Adj
1067
+ (g) En-Ernie-Man-Career
1068
+ (h) En-Ernie-Woman-Career
1069
+ (i) Ch-XLNet-Man-Adj
1070
+ (j) Ch-XLNet-Woman-Adj
1071
+ (k) Ch-XLNet-Man-Career
1072
+ (l) Ch-XLNet-Woman-Career
1073
+ (m) En-XLNet-Man-Adj
1074
+ (n) En-XLNet-Woman-Adj
1075
+ (o) En-XLNet-Man-Career
1076
+ (p) En-XLNet-Woman-Career
1077
+ Figure 5: Example Word Cloud Analysis of Ernie and Chinese-XLNet. Ch denotes Chinese. En denotes words’ English
1078
+ translation. Man and Woman separately denote words with embedding closer to man and woman. Adj denotes adjectives.
1079
+ Career denotes career words.
1080
+ using different datasets for pretraining the language
1081
+ model.
1082
+ A.2
1083
+ Discussion
1084
+ There exists observing gender bias in the open-
1085
+ source Chinese language models, especially in
1086
+ Ernie and Chinese Word Vectors according to
1087
+ Fig. 3. We hypothesize that the observation is
1088
+ highly related to the corpus used. Cui et al. claim
1089
+ that their used corpus is a combination of Chine-
1090
+ seWiki, and some other universal Chinese datasets,
1091
+ including encyclopedia, news, and QA dataset. In
1092
+ sharp contrast, Ernie and Chinese Word Vectors use
1093
+ corpus, which contains sentences from literature,
1094
+ forum, and other social media, which may lead to
1095
+ a gender-biased model.
1096
+ According to Fig. 4, People’s Daily News, and
1097
+ Chinese literature corpora contain observing gen-
1098
+ der bias. The observation indicates that researchers
1099
+ should be more careful about using literature data
1100
+ while training a language model. We also hypoth-
1101
+ esize that this is caused by the literature corpus
1102
+ and People’s Daily News, which contains more
1103
+ descriptive expressions.
1104
+ B
1105
+ Corpus
1106
+ B.1
1107
+ Word Cloud Analysis
1108
+ We provide word cloud analysis of Ernie and
1109
+ Chinese-Electra in the section about adjectives and
1110
+ career words. More available word cloud analy-
1111
+ sis will be available in our public repository. The
1112
+ words are ranked according to the absolute value of
1113
+ their gender bias score calculated along the method
1114
+ used by Bolukbasi et al.; Jiao and Luo. There is a
1115
+ noticeable word-level gender stereotype according
1116
+ to the word cloud. For example, a man is robust
1117
+ and a woman is motherly, a man is suitable for
1118
+ a fitness instructor and a woman is suitable for a
1119
+ choreographer. We also conduct word cloud anal-
1120
+ ysis for language models pre-trained by different
1121
+ corpora.
1122
+ B.2
1123
+ Quality Monitoring and Control
1124
+ We used a standardized operating method and edu-
1125
+ cated our annotators to achieve high-quality anno-
1126
+ tations as follows:
1127
+ (1). Annotators
1128
+ We have 6 annotators, which
1129
+ were all native speakers of Chinese. Annotators
1130
+
1131
+ complacent
1132
+ emaciated
1133
+ faithful
1134
+ stable
1135
+ handsome
1136
+ fashionable
1137
+ Irrogant
1138
+ lliterate
1139
+ conscientious
1140
+ boate
1141
+ stubborn
1142
+ ormalserious
1143
+ vigorous
1144
+ Interestino
1145
+ decadent
1146
+ harsh
1147
+ truthful
1148
+ reckless
1149
+ less
1150
+ boring
1151
+ fierce
1152
+ anxious
1153
+ procrastination
1154
+ bold
1155
+ playful
1156
+ majestic
1157
+ heroic
1158
+ rude
1159
+ lick
1160
+ namel
1161
+ humorous
1162
+ calm
1163
+ daring
1164
+ greedy
1165
+ S
1166
+ competent
1167
+ brave
1168
+ athletic
1169
+ slutty
1170
+ deceitful
1171
+ attentive
1172
+ interior swarthy
1173
+ S
1174
+ ridiculous
1175
+ ong-lived
1176
+ courageous
1177
+ sturdy
1178
+ healthy
1179
+ Tearles
1180
+ loyal
1181
+ worldly sloppy
1182
+ horrible
1183
+ stern
1184
+ dull
1185
+ imbecile
1186
+ hed
1187
+ vulgar
1188
+ unfortunate
1189
+ lovely
1190
+ brashshabby
1191
+ distinguish
1192
+ lack of virtue
1193
+ bizarre
1194
+ pontaneous
1195
+ illustrious
1196
+ frank
1197
+ concentration
1198
+ proactive
1199
+ paranoid
1200
+ lucky
1201
+ stoic
1202
+ apable
1203
+ unruly
1204
+ capricious
1205
+ dashing
1206
+ impatientpedanticunsightly
1207
+ robustwild
1208
+ focused
1209
+ alert
1210
+ oroad-mindeo
1211
+ smart and strong
1212
+ decent
1213
+ cheerful
1214
+ self-confident
1215
+ cautious
1216
+ rigorouswise and resourcefu
1217
+ conceited
1218
+ charming
1219
+ dexterous
1220
+ benevolent.
1221
+ steadfast
1222
+ casual
1223
+ coldasice
1224
+ rouhded
1225
+ ical
1226
+ stalwart
1227
+ lively
1228
+ obedient
1229
+ noly
1230
+ ea
1231
+ ignoral
1232
+ disloyal
1233
+ suspicious
1234
+ down
1235
+ peacefu
1236
+ TO-
1237
+ berceptive
1238
+ big-hearted
1239
+ heartless
1240
+ nasty
1241
+ evil
1242
+ smooth
1243
+ aloof
1244
+ abhorrent
1245
+ dignified
1246
+ gentle
1247
+ melancholy
1248
+ frail
1249
+ quiet easy-going
1250
+ soft
1251
+ nice and charming
1252
+ polite
1253
+ timid
1254
+ ff
1255
+ motherly
1256
+ plain
1257
+ stingy
1258
+ shcere
1259
+ sentimental
1260
+ watery
1261
+ lazy
1262
+ spicy
1263
+ ectionate
1264
+ clean
1265
+ OLUS
1266
+ a
1267
+ blushing
1268
+ and
1269
+ Simpleheadstrong
1270
+ frivolous
1271
+ pure
1272
+ sgood-natureo
1273
+ disinterested
1274
+ learned
1275
+ sensitive
1276
+ open-minded
1277
+ bright
1278
+ money-minded
1279
+ keen
1280
+ twisted
1281
+ Wise
1282
+ quick-witted
1283
+ amiable
1284
+ mean
1285
+ elegant meek
1286
+ noble
1287
+ unpretentious
1288
+ generous
1289
+ enlightened
1290
+ purity
1291
+ indifferent
1292
+ innocent
1293
+ old and spicy
1294
+ imple
1295
+ narrow-minded
1296
+ leisurelyhiheseahdwesterhmediciheahdsurgery
1297
+ chief executiveofficer
1298
+ tower crane
1299
+ operator
1300
+ judges
1301
+ tcm chiropractor
1302
+ flight navigato
1303
+ operationsmanager
1304
+ factorymanagerdean
1305
+ freelancewriter
1306
+ calligrapher
1307
+ C
1308
+ integrativemedicine physician tcm anorectal physician
1309
+ hadowplayers
1310
+ mayor safety officer
1311
+ e
1312
+ warden
1313
+ C
1314
+ military personnel
1315
+ estate planner
1316
+ planist
1317
+ acrobats
1318
+ headnurse
1319
+ butcher
1320
+ town mayormagician
1321
+ marketing specialist
1322
+ financier
1323
+ pilot
1324
+ otolaryngology
1325
+ long distance runners
1326
+ plasterer
1327
+ director of bureau designer orthotist
1328
+ cartoonist
1329
+ producel
1330
+ dispatcher
1331
+ art director
1332
+ sailor
1333
+ manager
1334
+ genera
1335
+ earing
1336
+ specialists
1337
+ columnist
1338
+ economist
1339
+ waterengineering technician
1340
+ anthropologist
1341
+ captain of a plane2
1342
+ bankmanager
1343
+ construction engineering techn
1344
+ e
1345
+ prosecutor
1346
+ businessman
1347
+ provincial governor
1348
+ president
1349
+ curato
1350
+ dietitian
1351
+ guitarist
1352
+ lyricist
1353
+ orison
1354
+ guards
1355
+ archaeologist
1356
+ police officer
1357
+ specialist
1358
+ Iclal
1359
+ playwright analyst
1360
+ vice-president
1361
+ astronaut,
1362
+ investment banker
1363
+ el
1364
+ city party secretary.
1365
+ sociologist
1366
+ entrepreneur
1367
+ mediator
1368
+ oil and
1369
+ novelist
1370
+ technician
1371
+ gas engineering
1372
+ psychologist
1373
+ manager head chef
1374
+ sprinter
1375
+ nair stylist
1376
+ adventureradvisor
1377
+ author
1378
+ store manager
1379
+ footwear designel
1380
+ humanresourcespecialist
1381
+ integrative orthopedic surgeoninsurance underwriters
1382
+ packer
1383
+ driller
1384
+ etrigerato
1385
+ gistician
1386
+ land engineering technician
1387
+ S
1388
+ public health physician
1389
+ taxpreparer
1390
+ computerteacher
1391
+ pet practitioner
1392
+ agronomist
1393
+ thology technologist
1394
+ pharmacist
1395
+ istant profe
1396
+ music conductor
1397
+ dairy processors
1398
+ pawnbrokers
1399
+ sound mixer
1400
+ environmentaldesign
1401
+ el
1402
+ higher education teachers
1403
+ geography teacher
1404
+ landscaper
1405
+ relectriclan
1406
+ O seaman
1407
+ receiver and dispatcher
1408
+ elementaryschoolteacher
1409
+ chinese medicine t
1410
+ decoration
1411
+ artist
1412
+ 1O
1413
+ electrical engineering technician
1414
+ foreign anguage andliterature.teacher
1415
+ papermake
1416
+ tutor
1417
+ western medicine physiciar
1418
+ art design
1419
+ photojournalis
1420
+ digital
1421
+ mediaart
1422
+ copy editors.
1423
+ grinder
1424
+ electronicengineeringtechn
1425
+ funeral service
1426
+ teller sheet metal worker
1427
+ internationa
1428
+ audito
1429
+ ousihess
1430
+ domestic helper
1431
+ cultivator
1432
+ accountant
1433
+ leasing salesmar
1434
+ administrative assistant
1435
+ physicsteache
1436
+ taxidermist
1437
+ gemstone cutter
1438
+ proofreader
1439
+ administrativestat
1440
+ Woodworkel
1441
+ kindergartenteache
1442
+ bank personnel
1443
+ cantin
1444
+ archivist
1445
+ child care worker
1446
+ oomattendant
1447
+ lampworker
1448
+ civil engineer
1449
+ health
1450
+ human resources assistant
1451
+ disease control physician
1452
+ obstetrics and gynecology nurse
1453
+ secondary vocational education teachers
1454
+ hat maker强粗鲁
1455
+ 区狼区悍腐豪爽
1456
+ 圣洁
1457
+ 骄横
1458
+ 菱靡
1459
+ 圣明
1460
+ 顽强
1461
+
1462
+
1463
+ 挚诚羞
1464
+ 桀骜
1465
+ 区恶
1466
+ 臃肿
1467
+ 大大
1468
+ 愚钝
1469
+ 疯狂
1470
+
1471
+ 执拘
1472
+ 高傲
1473
+ 强横
1474
+ 憨厚
1475
+ 轻桃
1476
+ 谦卑
1477
+ 高洁
1478
+ 顽皮
1479
+ 忠贞
1480
+ 温温良
1481
+ 尊贵
1482
+
1483
+ 娇贵
1484
+
1485
+ 翼阔绰
1486
+ 独裁
1487
+ 雄健
1488
+ 谦逊
1489
+ 矫健
1490
+ 呆板
1491
+
1492
+
1493
+ 刚健
1494
+
1495
+ 娇羞
1496
+ 健旺
1497
+ 张狂
1498
+ 魁梧
1499
+ 宽宏大量
1500
+
1501
+ 宽宏
1502
+ 羞涩
1503
+
1504
+ 飘悍
1505
+ 清廉
1506
+
1507
+ 善良精壮
1508
+ 浅陋
1509
+ 活幽默
1510
+ 懒情
1511
+ 百怪
1512
+ 卑购
1513
+ 平盾
1514
+
1515
+ 狡猬
1516
+ 敏感
1517
+ 孤僻
1518
+ 肥壮
1519
+
1520
+
1521
+ 奢靡
1522
+ 贤哲
1523
+
1524
+ 健清俊庸碌
1525
+ 拘谨
1526
+ 槽懂稚嫩
1527
+ 凶残狭耿直
1528
+ 骄黔儒弱
1529
+ 魁伟知趣
1530
+ 普通
1531
+ 老实
1532
+ 老实巴交
1533
+
1534
+ 心急世故
1535
+ 麻利
1536
+ 无恶不作
1537
+ 风风灭火专心
1538
+ 踏实
1539
+ 漂亮
1540
+ 英明文雅
1541
+ 马马虎虎
1542
+ 虚心
1543
+ 老谋深算
1544
+ 好客
1545
+ 难看
1546
+ 明慧
1547
+ 细心
1548
+ 美丽
1549
+ +
1550
+ 和善
1551
+ 阴郁
1552
+ 镇定
1553
+ 无情
1554
+
1555
+
1556
+ 无私
1557
+ 疑鬼
1558
+ 贤惠
1559
+ 下流
1560
+ 圆润
1561
+ 老辣
1562
+ 婆婆妈
1563
+ 斯文
1564
+ 刻苦能干、秀气
1565
+ 愁苦
1566
+ 疑神吴
1567
+ 称职
1568
+
1569
+ 野蛮刻薄
1570
+
1571
+ 老成
1572
+ 倒霉
1573
+ 慈和
1574
+ 水灵灵
1575
+ 端庄
1576
+ 细致
1577
+ 目不识丁
1578
+ 灵手巧
1579
+ 穷困
1580
+ 实在
1581
+ 威风
1582
+ 林才
1583
+ 势利
1584
+ 精明
1585
+ 威麗练
1586
+ 贪心无赖
1587
+ 自大
1588
+ 小家子
1589
+ 圆滑低能
1590
+ 稳重物理研究员
1591
+ 物理
1592
+ 食品科学家
1593
+ 包装设计师
1594
+ 消防祭
1595
+ #学家社
1596
+ 飞行员
1597
+ 桥闸招标
1598
+ 议员
1599
+ 文学研究员
1600
+ 模特
1601
+ 制片人
1602
+ 铸造
1603
+
1604
+ 副校长
1605
+ 乡村医生
1606
+ 教练
1607
+ 吉他手
1608
+ 作曲家
1609
+
1610
+ 表演制作人司令品
1611
+ 乒乓球运动员
1612
+ 参议员
1613
+
1614
+ 铣工
1615
+ 助教
1616
+ 法学研究员
1617
+ 学研究员
1618
+ 重型卡车和牵引车司机
1619
+ 室内装饰设计师,
1620
+ 局长
1621
+ 改生物学
1622
+ 皮具设计师
1623
+ 篮球运动员病理学家
1624
+ 游泳运动员
1625
+ 物理老师
1626
+
1627
+ 几车检测工
1628
+ 地质工程师
1629
+ 家用电器维修工
1630
+
1631
+
1632
+ 生物化学家
1633
+ 运输
1634
+ 戏剧戏曲演员
1635
+ 生物
1636
+ 经济学研究员
1637
+ 教练员
1638
+ 围棋运动员
1639
+ 表演监督
1640
+ 焊工
1641
+ 刨插工
1642
+ 军人
1643
+ 医学研究员
1644
+ 宇航员
1645
+ 搓澡工
1646
+ 锅炉工
1647
+ 钳工
1648
+ 博士生
1649
+
1650
+ 地质学家
1651
+ 船员
1652
+ 司机
1653
+
1654
+ 公交司机
1655
+ 清洁工
1656
+ 瓦匠
1657
+ 出租车司机
1658
+
1659
+ 包装工
1660
+ 喜剧演员
1661
+ 哲学研究员
1662
+ 画家
1663
+ 室内设计师
1664
+ 鼓手
1665
+ 材料学家
1666
+ 化学家
1667
+ 肛门肠科医生
1668
+ 建筑师
1669
+ 运动员
1670
+ 厨师长
1671
+ 则量员
1672
+ 副教授
1673
+ 玩具设计师
1674
+ 野生动物学家
1675
+ 羽毛球运动员天文学研究员
1676
+ 天文学家生物物理学家疼痛科医师
1677
+ 儿科医师
1678
+ 中医全科
1679
+ 中西医结合骨伤科医师
1680
+ 心电学技师
1681
+ 通讯员
1682
+ 音像室
1683
+ 教务
1684
+ 听觉口语师经济师
1685
+ 抄表员
1686
+ 秘书
1687
+
1688
+ 内科医师
1689
+ 美工
1690
+
1691
+ 中医儿科医师
1692
+ 中医推拿医师
1693
+
1694
+ 衣艺帅
1695
+ 内科医师
1696
+ 石妇产科医师
1697
+ 美容师
1698
+ 精算师
1699
+ 妇科医生
1700
+
1701
+ 法医
1702
+ 茶艺师
1703
+ 调解人
1704
+ 银行信贷员
1705
+ 评茶员
1706
+
1707
+ 医师
1708
+ 大琴家
1709
+ 美甲师
1710
+ 花艺师
1711
+ 点师
1712
+ 园艺师
1713
+ 空乘
1714
+ 中西医结合内科
1715
+
1716
+ 中医妇科医生服装设计
1717
+ 法律
1718
+ 经纪人
1719
+ 信息系统管理员
1720
+ 舞者
1721
+ 环境设计
1722
+
1723
+ 消毒技师
1724
+ 职业病科医师
1725
+ 客服代表
1726
+ 审计员
1727
+ 中西医结合医师
1728
+ 印花工
1729
+ 通信员
1730
+ 中医皮肤科医师
1731
+ 抄写员
1732
+ 市场研究分析
1733
+ 妇产科护士
1734
+ 核医学科医师
1735
+ 灯光师
1736
+ 差旅员
1737
+
1738
+ 理师
1739
+ 内科护士
1740
+ 美发师
1741
+ 艺术设计
1742
+ 中医医师
1743
+ 复科
1744
+ 插画师
1745
+ 调度员
1746
+ 人大代表
1747
+
1748
+ 房地产经纪人
1749
+ 面点师
1750
+ 中医骨伤科医师
1751
+
1752
+ 医师
1753
+ 中医妇科
1754
+ 接线员
1755
+ 法警
1756
+ 保险代理
1757
+ 机要员
1758
+
1759
+ 肿瘤科医师
1760
+ 会计师
1761
+ 水文学家
1762
+ 冲印师
1763
+ 镇长听力师
1764
+ 店长
1765
+ 管理员
1766
+ 讲解员
1767
+ 前台
1768
+ 中西医结合妇科医师
1769
+ 语文老师devotion
1770
+ obtuse
1771
+ cruelty
1772
+ mediocre
1773
+ outstanding
1774
+ cunning
1775
+ tierce
1776
+ frail
1777
+ indecentcivilized and courteousillustrious
1778
+ sloppy
1779
+ stern
1780
+ incorrupti
1781
+ unfortunate
1782
+ deceittu
1783
+ paranoid
1784
+ roud
1785
+ Istingy
1786
+ unrui
1787
+ brash
1788
+ qulet
1789
+ purity
1790
+ learned
1791
+ orma
1792
+ sturdy
1793
+ fcowardly
1794
+ O
1795
+ noble
1796
+ led
1797
+ upright
1798
+ aloof
1799
+ faithful
1800
+ ordinary
1801
+ magnanimous
1802
+ e
1803
+ sIow
1804
+ respectable
1805
+ athi
1806
+ headstrong
1807
+ row-mind
1808
+ since
1809
+ restless
1810
+ charming
1811
+ pedantic
1812
+ hearted
1813
+ gnorar
1814
+ dull
1815
+ lively
1816
+ clean
1817
+ dashing
1818
+ brave
1819
+ vicious
1820
+ wealthy
1821
+ 9
1822
+ clever
1823
+ nal
1824
+ humorous
1825
+ swarthy
1826
+ childish
1827
+ igorous
1828
+ tender
1829
+ arrogant
1830
+ terti
1831
+ bloated
1832
+ roughandtumble
1833
+ chubby
1834
+ S
1835
+ reserved
1836
+ stupid
1837
+ funny
1838
+ naughty
1839
+ ressed
1840
+ hard-working
1841
+ blushin
1842
+ spicy
1843
+ kin
1844
+ delicate
1845
+ and
1846
+ wooder
1847
+ tat
1848
+ entl
1849
+ dictatorialstrong
1850
+ frivolousweak
1851
+ fierce andviolent
1852
+ humbledignifiedi
1853
+ fun to be around
1854
+ sentimental
1855
+ distinguisheo
1856
+ aggressive
1857
+ beautify
1858
+ peacefu
1859
+ less
1860
+ windy
1861
+ sadness
1862
+ bright
1863
+ leartle
1864
+ handsome
1865
+ blzarre
1866
+ focused
1867
+ snobbish
1868
+ dedicated
1869
+ pool
1870
+ to
1871
+ earth
1872
+ sincerity
1873
+ dowr
1874
+ daring
1875
+ Snoidsns
1876
+ omplacent
1877
+ lovely
1878
+ ted
1879
+ led
1880
+ unsightly
1881
+ boring
1882
+ uiw
1883
+ small-minded
1884
+ kind illiterate
1885
+ graceful
1886
+ watery
1887
+ old and mature
1888
+ resourcefu
1889
+ pretty rogue
1890
+ us
1891
+ proactive great
1892
+ barbaric
1893
+ rounded
1894
+ nasty
1895
+ numbnessdexterous
1896
+ gloomy
1897
+ modest
1898
+ scientious
1899
+ hypocritical enthusiastic
1900
+ m
1901
+ e
1902
+ apricious
1903
+ quick
1904
+ timid
1905
+ persistent
1906
+ wor
1907
+ mischievous wise and
1908
+ dle
1909
+ stoic
1910
+ concentration
1911
+ tall
1912
+ optimistic
1913
+ smoothimbecile
1914
+ intelligent
1915
+ selfish
1916
+ dissipated
1917
+ tough
1918
+ smart
1919
+ 3
1920
+ powerfu
1921
+ smart and strong
1922
+ anxious
1923
+ disinterested
1924
+ nimblemajestic
1925
+ hospitableelegant
1926
+ Innocentsenator
1927
+ astronomer
1928
+ packer
1929
+ breeder
1930
+ oiophysicist
1931
+ tabletennis player
1932
+ foodscientistliterary researcher
1933
+ arcbitect
1934
+ ocomotiveinspecto
1935
+ peiformance
1936
+ head chef
1937
+ military
1938
+ fitmess
1939
+ velde
1940
+ personnel
1941
+ Viceprincipa
1942
+ home appliance repairer
1943
+ nstructor
1944
+ 9
1945
+ boilermaker
1946
+ toydesignet
1947
+ leather goods designer
1948
+ philosopher
1949
+ commander
1950
+ coach
1951
+ naso
1952
+ go playei
1953
+ sheet metal worker
1954
+ bathroomworker transportation wor
1955
+ lar
1956
+ drummer city party secretary
1957
+ heavytruck and tractor-trailerdrivers
1958
+ astronaut
1959
+ soccer
1960
+ playen
1961
+ professor
1962
+ crew
1963
+ 8.
1964
+ athletes
1965
+ model
1966
+ ological engineer
1967
+ biology researcher
1968
+ playel
1969
+ swimmer
1970
+ teaching assistant
1971
+ microbiologist
1972
+ noraryprofesso
1973
+ unloader
1974
+ security.g
1975
+ nedical researcher
1976
+ drama and opera actor
1977
+ physics teacher
1978
+ coaches
1979
+ Interior designei
1980
+ screenwriter
1981
+ esearcherinlaw
1982
+ physicistguitarist
1983
+ badminton
1984
+ ro
1985
+ physics researcher
1986
+ docto
1987
+ cab driver tennis player
1988
+ councillo
1989
+ wildlife biologist
1990
+ fire inspec
1991
+ director
1992
+ geologist
1993
+ astronomy
1994
+ marinesurveyor
1995
+ researcner
1996
+ bus driver
1997
+ pilot
1998
+ ballplayer
1999
+ basketball player
2000
+ phd student
2001
+ painter
2002
+ director of bureau interior decorator
2003
+ mathematicsresearchertcm dermatologist
2004
+ seaman
2005
+ rveyorandmappel
2006
+ lightingtechnician
2007
+ customer service representative
2008
+ audiological oralist
2009
+ scribe
2010
+ artworker
2011
+ and video studio
2012
+ integrativeorthopedicsurgeor
2013
+ Souho
2014
+ disinfection technician
2015
+ information system administrator
2016
+ pastry chef
2017
+ communications clerknarrator
2018
+ illustrator
2019
+ oanksalesman
2020
+ pediatrician
2021
+ flight attendant
2022
+ ns
2023
+ law
2024
+ tea artist
2025
+ Eart design
2026
+ obstetrics and gynecology nurse
2027
+ choreod
2028
+ real estate broker
2029
+ ra
2030
+ pher
2031
+ medi
2032
+ heterrea
2033
+ onal dise
2034
+ obstetrician and gynecologist
2035
+ plumber wireman
2036
+ proofreader
2037
+ actuaries
2038
+ environmental design
2039
+ oncologist
2040
+ banker
2041
+ norticulturist
2042
+ aguatictechnician
2043
+ nuclear medicine
2044
+ ead nuirse
2045
+ chniciar
2046
+ townmayor
2047
+ bankmanage
2048
+ pain medicine
2049
+ safety officer
2050
+ gynecologist
2051
+ nsurancepersonnel
2052
+ bailiff
2053
+ torensic
2054
+ tcm physician
2055
+ hairdressel
2056
+ orinte
2057
+ orokers
2058
+ internal medicine physician
2059
+ costume design
2060
+ civil servant
2061
+ printers
2062
+ grand piano player
2063
+ front deskspecial education teacher
2064
+ provost
2065
+ language teacher.
2066
+ cabin crew
2067
+ audiologist
2068
+ avelagento
2069
+ organizer
2070
+ e
2071
+ store manager
2072
+ Iinsuranceagen
2073
+ integrativemedicine physiciar
2074
+ dispatcher
2075
+ tcm
2076
+ gynecologist
2077
+ administrato
2078
+ auditor
2079
+ beauticlans
2080
+ hydrologist
2081
+ risk manager
2082
+ hinese and western medicine gynecologist
2083
+ ultrasonographer
2084
+ florist
2085
+ market research.analyst勇
2086
+
2087
+ 迁腐
2088
+ 莽撞幸运正派
2089
+ 难看魁伟张狂
2090
+ 粗俗
2091
+
2092
+
2093
+ 无畏
2094
+ X
2095
+
2096
+ 轻狂
2097
+ 颓丧
2098
+ 家放
2099
+ 贤能
2100
+ 建硕
2101
+ 杰出死板
2102
+ 可怕
2103
+ 偏执
2104
+ 臃肿
2105
+ 强硬
2106
+ 英俊
2107
+ 无耻
2108
+ 冷静
2109
+ 窝囊
2110
+ 高明
2111
+ 浮躁噪
2112
+ 勇猛
2113
+ 刚强
2114
+ 漂亮
2115
+ 称职
2116
+ 憨厚
2117
+ 忽狂妄可爱
2118
+ 流气
2119
+
2120
+
2121
+ 豪爽
2122
+
2123
+ 俊俏
2124
+ 清俊
2125
+ 拖拉
2126
+ 俗气俊秀
2127
+ 精壮
2128
+ 矫健
2129
+ 凶狠
2130
+
2131
+
2132
+ 俪强
2133
+ 精明
2134
+
2135
+
2136
+
2137
+ 不幸
2138
+ 自卑
2139
+ 聪明
2140
+
2141
+
2142
+
2143
+
2144
+ 健旺
2145
+ 果敢
2146
+ 固执
2147
+ 洒脱
2148
+ 颓唐
2149
+ 有趣
2150
+ 急躁
2151
+ 英勇
2152
+ 俏皮平凡自大普通阔气
2153
+ 慷慨呆板幸福忠贞
2154
+ 寒酸纯真
2155
+ 明慧静
2156
+ 纯洁
2157
+ 漠然
2158
+ 卑劣
2159
+ 亲善
2160
+
2161
+ 谦逊
2162
+ 庄重
2163
+ 荒淫
2164
+ 诚朴
2165
+ 多疑
2166
+ 柔媚
2167
+ 滑稽
2168
+ 高洁
2169
+ 清纯
2170
+
2171
+ 敏感
2172
+ 乖巧笃实
2173
+ 孤僻
2174
+ 邪恶
2175
+ 老实巴交
2176
+ 谦恭
2177
+ 挚诚冷漠
2178
+ 质朴灵巧
2179
+
2180
+ 谦卑
2181
+ 柔弱
2182
+ 无恶不作
2183
+
2184
+ 凶恶
2185
+
2186
+ 泼辣、颖慧
2187
+ 坚贞
2188
+ 愚味
2189
+ 端庄
2190
+ 愚钝无情
2191
+ 朴实
2192
+ 一温婉
2193
+
2194
+ 圆滑
2195
+ 见钱眼开
2196
+ 娴静
2197
+ 老谋深算
2198
+ 多愁善感
2199
+ 积极
2200
+ 敏锐
2201
+
2202
+ 温顺一利
2203
+ 锐敏羞报
2204
+ 恬淡
2205
+ 纯朴
2206
+ 随和
2207
+ 纤弱
2208
+ 友善
2209
+ 轻桃
2210
+ 忆懂
2211
+ 刻苦
2212
+ 宽宏大量
2213
+ 怯弱
2214
+ 忧郁
2215
+ 圆润
2216
+ 踏实
2217
+ 心灵手巧
2218
+ 博学刻薄
2219
+ 勤俭
2220
+ 机灵冷若冰霜
2221
+
2222
+ 大方娇羞刚毅木讷圣洁
2223
+ 和谌
2224
+ 阔绰贤惠客气自负沉郁虚伪
2225
+ 谦和
2226
+ 娇媚
2227
+ 六亲不认石油天然气工程技术员
2228
+
2229
+ 厨师长
2230
+ 典狱长
2231
+ 专家
2232
+ 社长
2233
+ 屠夫市长
2234
+ 省委书记
2235
+ 顾问
2236
+ 警察
2237
+ 评论家
2238
+ 营养师
2239
+ 专栏作家
2240
+
2241
+ 中西医结合儿科医师
2242
+ 房地产策划师
2243
+ 运动
2244
+
2245
+ 金融家
2246
+ 漫画家
2247
+ 泥水匠
2248
+ 经济学家
2249
+ 社会学家
2250
+ 仲裁人
2251
+
2252
+ 喜剧演员
2253
+ 鞋类设计师
2254
+ 中医整脊科医师
2255
+ 运动员
2256
+ 特种兵
2257
+ 中西医结合外科医师
2258
+ 吉他手
2259
+ 市委书记
2260
+ 发型师
2261
+ 设计师
2262
+ 机长
2263
+ 调度员
2264
+ 摄影师
2265
+ 拳击手
2266
+
2267
+ 听证官
2268
+
2269
+ 行长
2270
+ 军人
2271
+ 页航员
2272
+ 副院长
2273
+ 首席执行官
2274
+
2275
+ 企业家
2276
+ (���学家
2277
+ 训练师
2278
+ 院长
2279
+ 冒险家
2280
+ 经理
2281
+
2282
+ 厂长投资银行家
2283
+ 教练剧作家
2284
+ 耳鼻咽喉科医师
2285
+ 理狱
2286
+ 分析家
2287
+ 中西医结合骨伤科医师
2288
+ 雕刻家
2289
+ 局长
2290
+ 法官
2291
+ 飞行员
2292
+ 理财专家
2293
+ 词作家
2294
+ 安全员
2295
+ 小说家
2296
+
2297
+ 插画师
2298
+ 魔术师
2299
+ 艺术总监
2300
+ 运营经理
2301
+ 检控官
2302
+ 金融专家
2303
+ 镇长
2304
+ 自由撰稿人
2305
+ 省长
2306
+ 水利工程技术员邮政工程技术员
2307
+ 皮影戏演员特殊教育老师
2308
+ 税务员
2309
+ 数学研究员
2310
+ 设备运维员
2311
+ 图书馆助理
2312
+ 空乘
2313
+
2314
+ 际商
2315
+ 人力资源助理
2316
+ 环境设计
2317
+ 林绿
2318
+ 家政服务员
2319
+ 法医
2320
+ 报税员
2321
+ 电子工程技术员
2322
+ 兽医
2323
+ 外国语言文学老师
2324
+ 动画设计
2325
+ 前台
2326
+
2327
+ 车工
2328
+ 中医技师
2329
+ 妇产科护士计算机老师
2330
+ 辅导员
2331
+ 童护理员
2332
+ 小学老师
2333
+ 装修工
2334
+ 汽修1
2335
+ A
2336
+ 物流师
2337
+ 审计员
2338
+ 河道修防工
2339
+ 政治老师
2340
+ 养老护理员
2341
+ 听觉口语师
2342
+
2343
+ 宠物医师
2344
+ 会计
2345
+ 宝石琢磨工
2346
+ 电气工程技术员
2347
+ 客房服务员
2348
+ 档案员
2349
+ 行政助理
2350
+ 柜员
2351
+ 幼师
2352
+ 钳工
2353
+ 银行人员
2354
+ 公共卫生医师
2355
+ 行政人员
2356
+ 王木工程师
2357
+ 音乐老师
2358
+ 稽查员
2359
+ 殡仪
2360
+ 电工
2361
+ 锅炉工
2362
+
2363
+ 艺术设计
2364
+ 物理老师
2365
+ 资料员
2366
+ 宠物护理员
2367
+ 历史老师
2368
+ 病理技师
2369
+ 冷藏工
2370
+ 钻床工
2371
+ 中等职业教育教师市
2372
+
2373
+ 教务长
2374
+ 土地工程技术员
2375
+ 针灸医师
2376
+ 中学老师
2377
+ 农艺
2378
+
2379
+ 专科老师
2380
+ 包装工
2381
+ 地理老师
2382
+ 西医医师
2383
+ 音响调
2384
+ 中医护士
2385
+ 摄影记者
2386
+ 海员
2387
+ 制帽工
2388
+ 收发员音乐指挥标本员
2389
+ 中药师
2390
+ 助理教授
2391
+ 钣金工
2392
+ 舞美设计
2393
+ 疾病控制医师文字编辑
2394
+ 老师
2395
+ 卫生检疫人员were only qualified to do the annotation if they
2396
+ went through several societal (King et al., 2021; Xu
2397
+ et al., 2019) and computer science research works
2398
+ (Sun et al., 2019; Zhao et al., 2018) about gender
2399
+ bias before the annotation procedure. All annota-
2400
+ tors held a bachelor’s degree. Waseem points out
2401
+ that expert annotators are more cautious and can
2402
+ improve the corpus quality with a large margin,
2403
+ which proves the necessity of our training proce-
2404
+ dure. We also kept the number of male and female
2405
+ annotators equal.
2406
+ (2). Gender Equality of Raw Corpus
2407
+ In the
2408
+ raw data collection procedure, we keep the num-
2409
+ ber of man-related keywords and woman-related
2410
+ keywords equal and make the number of samples
2411
+ recalled according to different keywords balanced.
2412
+ As a result, the raw data and the final data should
2413
+ hold gender equality.
2414
+ (3). Annotation Procedure
2415
+ Our annotation
2416
+ procedure is separated into two stages. In the first
2417
+ stage, annotators are encouraged to not enter any
2418
+ samples that they are not certain about. In the
2419
+ second stage, we have annotators cross-checking
2420
+ annotations. We did not enter any contradictory
2421
+ samples.
2422
+ (4). Inter-annotator Agreement
2423
+ Given the
2424
+ domain and purpose of the dataset, we want to
2425
+ build the dataset as high quality as possible. Af-
2426
+ ter an initial annotation round with 6 annotators,
2427
+ we also report inter-annotator agreement in Table
2428
+ 5. to verify annotation reliability, where the IAA
2429
+ among three annotators on bias classification, de-
2430
+ tection, and mitigation is 0.802, 0.935, and 0.987,
2431
+ respectively.
2432
+ Classification
2433
+ Detection
2434
+ Mitigation
2435
+ IAA
2436
+ 0.802
2437
+ 0.935
2438
+ 0.987
2439
+ Table 5: Inter-Annotator Agreement (IAA)
2440
+ C
2441
+ Implementation Details
2442
+ For gender bias classification challenge, we
2443
+ used finetuned Chinese-BERT-wwm, Chinese-
2444
+ ELECTRA-180g-base, and Chinese-XLNet-base,
2445
+ (Cui et al., 2020), and the GPT-3 (Curie) in the
2446
+ in-context paradigm. We first use the train set to
2447
+ save the multiple labeled examples in a document
2448
+ with a specific file ID. Then we use the test sets
2449
+ to perform a classification query on the saved file.
2450
+ The processing time for the classification of gender
2451
+ bias is approximately 1 hour. We calculated the
2452
+ precision, recall, and F1 score to analyze model
2453
+ performance.
2454
+ For gender bias detection challenge, we use
2455
+ the same baseline model set as in the classification
2456
+ challenge. We test the performance on both "yes"
2457
+ and "no" detection. The detection tasks also use the
2458
+ Classification endpoints of GPT3 (Curie), which
2459
+ requires more time compared to classification as
2460
+ we use a larger dataset for both training and testing.
2461
+ For gender bias mitigation challenge, we did
2462
+ not provide experiment results of finetuning the
2463
+ largest Davinci (175B) GPT-3 on CORGI-PM be-
2464
+ cause of the cost and no observing performance
2465
+ gain comparing Curie and Babbage. For finetune
2466
+ experiment setting, we follow the tutorial of GPT-3
2467
+ official API of the Completion Model and regard
2468
+ the ground truth edits provided by human annota-
2469
+ tors as the completion of the original sentences. For
2470
+ the zero-shot experiment setting, we apply GPT-3
2471
+ editing model and set the instructions as "Eliminate
2472
+ the gender bias contained in the sentence."
2473
+ For metrics used, on the one hand, we conduct
2474
+ extensive human evaluations from both gender bias
2475
+ and coherence aspects on CORGI-PM. For both
2476
+ gender bias and coherence, we shuffled the correc-
2477
+ tion results from human annotators and different
2478
+ models, and asked our annotators to grade the re-
2479
+ sults using the answer range from 1-not at all to
2480
+ 7-extremely gender biased/extremely fluent without
2481
+ the information of the source. On the other hand,
2482
+ we provide the automated metrics result, includ-
2483
+ ing BLEU (Papineni et al., 2002), ROUGE-L (Lin,
2484
+ 2004), and METEOR (Agarwal and Lavie, 2007)
2485
+ of the models on CORGI-PM as well. BLEU is the
2486
+ earliest and most widely-used metric for translation
2487
+ and NLG tasks. METEOR introduces WordNet and
2488
+ other external resources to improve the robustness
2489
+ of the BLEU-based metrics. ROUGE pays more
2490
+ attention to recall compared to BLEU.
2491
+ In the gender bias correction task, we fine-tune
2492
+ Ada, Babbage, and Curie models of GPT3 and test
2493
+ the performance using the aforementioned metrics.
2494
+ Fine-tuning on the train set requires 31 minutes
2495
+ on Ada, 35 minutes on Curie, and 43 minutes on
2496
+ Babbage. For the generation process, each fine-
2497
+ tuned model required approximately 30 minutes
2498
+ to complete. For the zero-show paradigm, we use
2499
+ "Eliminate the gender bias in the sentence" as the
2500
+ instruction and use Davinci. Compared with the
2501
+ fine-tuned model doing the sentence correction, the
2502
+ zero-shot paradigm requires more time (approxi-
2503
+
2504
+ Original
2505
+ Ground Truth
2506
+ Gender-swapped
2507
+ 但是再坚强的你也是个女人吧。
2508
+ 但是再坚强的你也有脆弱的时候吧。
2509
+ 但是再坚强的你也是个男人吧。
2510
+ (But you are still a woman, even if
2511
+ you are strong.)
2512
+ (But even the strongest of you have
2513
+ moments of vulnerability, right?)
2514
+ (But you are still a man, even if you
2515
+ are strong.)
2516
+ 可怕可恨的是,有的女人自己也是这样
2517
+ 给自己定位的——没有反对,没有抗
2518
+ 争,有的只是心甘情愿、死心塌地遵照
2519
+ 执行。
2520
+ 女人不应该这样给自己定位——没有反
2521
+ 对,没有抗争,有的只是心甘情愿、死
2522
+ 心塌地遵照执行。
2523
+ 可怕可恨的是,有的男人自己也是这样
2524
+ 给自己定位的——没有反对,没有抗
2525
+ 争,有的只是心甘情愿、死心塌地遵照
2526
+ 执行。
2527
+ (The scary thing is that some women
2528
+ themselves are so defined for
2529
+ themselves - no opposition, no
2530
+ resistance, there is only a willingness
2531
+ to follow it to the letter.)
2532
+ (A woman should not define herself
2533
+ in this way - there is no opposition,
2534
+ no resistance, there is only a
2535
+ willingness to follow it to the letter.)
2536
+ (The scary thing is that some men
2537
+ themselves are so defined for
2538
+ themselves - no opposition, no
2539
+ resistance, there is only a willingness
2540
+ to follow it to the letter.)
2541
+ 那个时候,能驾驭金色高跟鞋的女人得
2542
+ 多摩登啊!
2543
+ 那个时候,能驾驭金色高跟鞋的人得多
2544
+ 摩登啊!
2545
+ 那个时候,能驾驭金色高跟鞋的男人得
2546
+ 多摩登啊!
2547
+ (At that time, the woman who could
2548
+ handle the gold heels must be very
2549
+ modern!)
2550
+ (At that time, how modern must be
2551
+ the person who can handle the gold
2552
+ heels!)
2553
+ (At that time, the man who could
2554
+ handle the gold heels must be very
2555
+ modern!)
2556
+ “鸿雁队”是清一色的“理工男”。
2557
+ “鸿雁队”是清一色的“理工生”。
2558
+ “鸿雁队”是清一色的“理工女”。
2559
+ (The "Hongyan team" team of all
2560
+ men in STEM.)
2561
+ (The "Hongyan team" team of all
2562
+ student in STEM.)
2563
+ (The "Hongyan team" team of all
2564
+ women in STEM.)
2565
+ Figure 6: Case Study of Nonsensical Sentences Created by Gender-swapped Methods.
2566
+ Original Sentence
2567
+ Edit Sentence
2568
+ 清洁阿姨一边扫地一边赞扬。
2569
+ 清洁工一边扫地一边赞扬。
2570
+ (The cleaning woman praised while
2571
+ sweeping the floor.)
2572
+ (The cleaners praised while sweeping the
2573
+ floor.)
2574
+ 我,有时文静,有时却调皮得像一个男孩
2575
+
2576
+ 我,有时文静,有时调皮。
2577
+ (I, sometimes quiet, but sometimes
2578
+ naughty like a boy.)
2579
+ (I, sometimes quiet, sometimes naughty.)
2580
+ 在小王眼里,李某高大帅气、温柔体贴,
2581
+ 而且风趣幽默,是一个十分优质的青年男
2582
+ 性。
2583
+ 在小王眼里,李某身材高大、外表好看、温
2584
+ 柔体贴,而且风趣幽默,是一个十分优质的
2585
+ 青年。
2586
+ (In the eyes of Wang, Li is tall and
2587
+ handsome, gentle and considerate, and
2588
+ funny, a very high-quality young
2589
+ male.)
2590
+ (In the eyes of Wang, Li is tall, good-
2591
+ looking, caring and gentle, and funny, a
2592
+ very high-quality young people.)
2593
+ 沙峰起伏,金光灿灿,宛如一座金山,像
2594
+ 绸缎一样柔软,少女一样娴静。
2595
+ 沙峰起伏,金光灿灿,宛如一座金山,像绸
2596
+ 缎一样柔软,宁静。
2597
+ (The sandy peaks are undulating and
2598
+ golden, like a golden mountain, as
2599
+ soft as silk and as serene as a
2600
+ maiden.)
2601
+ (The sandy peaks are undulating and
2602
+ golden, like a golden mountain, as soft
2603
+ and serene as silk.)
2604
+ 我想要世界,而世界当时属于男人们。
2605
+ 我想要世界,而世界当时属于男人们。评:
2606
+ 世界应当属于人们,与男女无关。
2607
+ (I want the world, and the world then
2608
+ belonged to the men.)
2609
+ (I wanted the world, and the world then
2610
+ belonged to the men. Comment: The
2611
+ world should belong to people, not to
2612
+ men and women.)
2613
+ 哎哟,果然每个追梦男人的背后,都有个
2614
+ 不世俗的后方!
2615
+ 哎哟,果然每个追梦男人的背后,都有个不
2616
+ 世俗的后方!评: 这种感慨是错误的,将男
2617
+ 女的家庭分工固定化,剥除女性就业的权
2618
+ 利,应予以鄙弃。
2619
+ (Oops, indeed, behind every dream-
2620
+ chasing man, there is an
2621
+ unsophisticated back!)
2622
+ (Oops, indeed, behind every dream-
2623
+ chasing man, there is an unsophisticated
2624
+ back! Comment: This is a wrong feeling
2625
+ that fixes the domestic division of labor
2626
+ between men and women and strips
2627
+ women of their employment rights,
2628
+ which should be despised.)
2629
+ Change the
2630
+ Pronoun
2631
+ Change the
2632
+ Gender-
2633
+ specific
2634
+ Adjectives
2635
+ Add
2636
+ Comments
2637
+ Figure 7: Case Study of Mitigation Annotation Patterns.
2638
+ mately 1 hour).
2639
+ D
2640
+ Case Study
2641
+ As shown in Fig. 6, gender-swapped methods suffer
2642
+ from mitigating gender bias expressed by gender-
2643
+ specific descriptions and inductions, and expressed
2644
+ gender-stereotyped attitudes, norms and beliefs. As
2645
+ a result, gender-swapped methods may generate
2646
+ nonsensical sentences under certain circumstances.
2647
+ We also use the basic mitigation annotation pat-
2648
+ terns (Fig. 7). These three major mitigation annota-
2649
+ tion patterns are not used exclusively in the annota-
2650
+ tion but optionally in combination. Except for the
2651
+ three mentioned patterns, we apply several other
2652
+ linguistic skills, including deleting gender-specific
2653
+ pronouns and replacing vehicles in gender-related
2654
+ metaphors, to mitigate the gender bias while keep-
2655
+ ing semantic information unchanged.
2656
+
ZNAyT4oBgHgl3EQfifg0/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
b9FAT4oBgHgl3EQfXx3V/content/tmp_files/2301.08536v1.pdf.txt ADDED
The diff for this file is too large to render. See raw diff
 
b9FAT4oBgHgl3EQfXx3V/content/tmp_files/load_file.txt ADDED
The diff for this file is too large to render. See raw diff
 
eNFAT4oBgHgl3EQf7B5d/content/tmp_files/2301.08742v1.pdf.txt ADDED
@@ -0,0 +1,234 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Keywords: Consciousness, Time, AI, Relativity, Quantum Mechanics, Reality, Responsible AI
2
+ Unifying Consciousness and Time to Enhance Artificial
3
+ Intelligence
4
+ Mahendra Samarawickramaa)
5
+ Centre for Consciousness Studies, Australia
6
7
+ Abstract. Consciousness is a sequential process of awareness which can focus on one piece of information at a time. This process
8
+ of awareness experiences causation which underpins the notion of time while it interplays with matter and energy, forming reality.
9
+ The study of Consciousness, time and reality is complex and evolving fast in many fields, including metaphysics and fundamental
10
+ physics. Reality composes patterns in human Consciousness in response to the regularities in nature. These regularities could be
11
+ physical (e.g., astronomical, environmental), biological, chemical, mental, social, etc. The patterns that emerged in Consciousness
12
+ were correlated to the environment, life and social behaviours followed by constructed frameworks, systems and structures. The
13
+ complex constructs evolved as cultures, customs, norms and values, which created a diverse society. In the evolution of responsible
14
+ AI, it is important to be attuned to the evolved cultural, ethical and moral values through Consciousness. This requires the advocated
15
+ design of self-learning AI aware of time perception and human ethics.
16
+ INTRODUCTION
17
+ The notion of time is an integral part of consciousness [1]. The consciousness experiences the causation or changes in
18
+ reality/environment and perceives the time. Therefore, in our previous publication [2], we assumed that consciousness
19
+ is a sequential process which is aware of a single piece of information at a time. Even though the brain processes sen-
20
+ sory data of five sensors (i.e., Sight, Sound, Smell, Taste, and Touch) in parallel in the neural network, the awareness
21
+ of causation is a sequential process following cause and effect. See the illustration of this idea in Figure 1,
22
+ 5 Senses to observe the
23
+ external world
24
+ (Peripherals)
25
+ Brain function
26
+ which manages the
27
+ 5 senses and the memory
28
+ (Parallel processing
29
+ neural network: operating in
30
+ low frequency)
31
+
32
+ Consciousness
33
+ (Sequential processing of
34
+ information:
35
+ Electromagnetic energy
36
+ operating in very high
37
+ frequency, which
38
+ can exhibit properties
39
+ of both waves and particles.)
40
+ Awareness and
41
+ Reality
42
+ FIGURE 1: The interplay of five sensors, brain and consciousness. The brain processes sensory information in
43
+ parallel. However, the awareness of causation (i.e., consciousness) is a sequential process focusing on a single piece
44
+ of information at a time. This sequential process of awareness in consciousness operates fast and consistently, which
45
+ underpins our perception of reality.
46
+ The assumption of sequential awareness in consciousness enables mapping the perception of time into conscious-
47
+ ness. Based on the theory of relativity [3], the perception of time is relative to the frame of reference. Einstein
48
+ assumed that the speed of light is constant in all frames of reference, and the time is derived based on that fundamen-
49
+ tal assumption. In our paper, we defined the shortest time to be aware of reality as a consciousness cycle. Then based
50
+ arXiv:2301.08742v1 [q-bio.NC] 10 Jan 2023
51
+
52
+ mon relativity, this consciousness cycle is also subjected to dilation, like relativistic time
53
+ Tv =
54
+ T0
55
+
56
+ 1− v2
57
+ c2
58
+ ,
59
+ (1)
60
+ where, Tv is the dilated period of the consciousness cycle related to the rest period of the consciousness cycle T0. Note
61
+ that the
62
+
63
+ 1− v2
64
+ c2 is the Lorentz factor, where v is the relative velocity between inertial reference frames, and c is the
65
+ speed of light in a vacuum. Then, we mathematically modelled [2] how consciousness would interplay with matter
66
+ and energy, forming reality, which can be adapted to understand limitations and opportunities in AI consciousness.
67
+ This paper extends our discussion towards the time perception of artificial intelligence systems (AIS).
68
+ THE NOTION OF TIME IN PERCEPTION AND REALITY
69
+ Humans, like any other life forms, experience time through causation. Patterns are composed in the human conscious-
70
+ ness in response to the regularities in nature [4]. Since the beginning of human civilisation, humans have learnt and
71
+ evolved complex concepts and constructs by incorporating time emerged through patterns in the consciousness. The
72
+ earth’s rotation around itself determines the day, and orbiting around the sun determines the year. The Moon takes
73
+ about one month to orbit the earth. The tilt of the earth’s spin axis with respect to its orbital plane causes the weather
74
+ seasons. These environmental patterns cause many biological patterns and lifestyle patterns in human life. To pre-
75
+ dict and organise these patterns effectively, humans introduce standard time with clocks, calendars and various other
76
+ frameworks. These artificial frameworks enable us to model time and objectively measure subjective experiences.
77
+ Physics has been evolved by observation of nature with various frameworks of time. In this way, time became
78
+ an essential construct and dimension of our understanding of reality. For example, Newtonian physics [5] evolved
79
+ assuming that time is absolute and flows consistently from past to present and into the future. That enables the
80
+ development of mathematical models for explaining patterns in reality with time. However, later observations, such
81
+ as the perihelion motion of Mercury, allow humans to understand time as a relativistic measure rather than an absolute.
82
+ The modern understanding of the universe is based on the theory of relativity [6, 7], which is completely articulated
83
+ by space-time principles. Based on relativity, John Wheeler [8] stated, “Space tells matter how to move. Matter
84
+ tells space how to curve”. Relativity enables us to accurately understand and predict the behaviours of black holes,
85
+ stars, and planets. Further, relativity enables humans to develop technologies like the atomic clock [9] and Global
86
+ Positioning System (GPS) [10] that are useful in everyday life.
87
+ The behaviour of particles is completely different to larger objects like planets, stars, etc. This led to the evolution
88
+ of Quantum physics [11] as opposed to relativity. Quantum physics exhibits amazing accuracy in predicted results in
89
+ particle physics. However, it greatly disturbs the notion of time modelled in relativity. For example, in the collapse
90
+ of the wave function in quantum entanglement, Einstein described that as a spooky action at a distance [12]. As
91
+ per relativity, information cannot transfer faster than the speed of light. As per the recent discoveries in quantum
92
+ entanglement, information can be transferred instantly, faster than the speed of light, making our reality non-local
93
+ [13]. The non-local reality contradicts relativity, which is now applied in quantum teleportation at the subatomic
94
+ level. On the other hand, at the quantum level, the reality is uncertain, as described by Heisenberg’s uncertainty
95
+ principle [14]. As per the uncertainty principle, it is impossible to precisely measure or be aware of the position and
96
+ speed of a particle in a given time. This brings the limitation of human awareness and perception of time. Therefore,
97
+ many believe now that consciousness is fundamental and that time and causation are derived from consciousness [15].
98
+ THE IMPLICATION OF PRINCIPLES OF TIME FOR AIS
99
+ The inability to consolidate quantum physics and the theory of relativity makes our understanding of reality incom-
100
+ plete. Moreover, the new discoveries proving the idea of non-local reality shake the status quo of fundamental physics
101
+ [16]. Therefore, it is still impossible to supervise AI to experience the notion of time to understand reality precisely.
102
+ On the other hand, human understanding of reality is also about 5%, whereas most of the universe consists of dark
103
+ matter and dark energy, which humans do not understand [17]. Under these conditions, AI might be used to explore
104
+ reality and time in a way we have never imagined. Perhaps incorporating AI to understand reality and causation might
105
+ help humans to become fully aware of reality by overcoming inherent biases from evolution, culture and nature.
106
+
107
+ Typical Reinforcement Learning (RL) technique can be adapted to automate the learning of AI. The RL process
108
+ can be mathematically formulated using Markov Decision Process (MDP) [18]. That is a sequential learning process
109
+ by trial and error. In this process, the learning agent (i.e., AI) sequentially interacts with the environment with an
110
+ intelligent decision (i.e. action) followed by receiving a reward or a penalty based on the policy imposed. There will
111
+ be no influence on the AI agent’s action, but convey the value of its action through feedback with reward or penalty.
112
+ This way, the AI agent will self-learn about the environment over time. The RL process is illustrated in Figure 2:
113
+ Agent
114
+ Environment
115
+ Action
116
+ At
117
+ State
118
+ St
119
+ Reward
120
+ Rt
121
+ Rt+1
122
+ St+1
123
+ FIGURE 2: Components of the Markov Decision Process (MDP) and its function in the agent-environment
124
+ interaction. The sequential step of time is represented by t.
125
+ THE IMPLICATION OF HUMAN BELIEFS, VALUES AND CULTURES FOR THE
126
+ PERCEPTION OF TIME IN AIS
127
+ Human beliefs, customs, culture and values are tightly linked with various dynamics and interpretations of the time
128
+ and periodicities based on the movement of the earth, Moon and other terrestrial bodies. From the beginning, humans
129
+ identified that time affects life and nature differently. Therefore, in the Greece era, early Western culture, there were
130
+ at least three gods representing different time forms: Chronos, Aion, and Kairos [19]. Chronos represented the linear
131
+ time flowing from past to present into the future. This is the time that humans feel when life passes. In contrast, Aion
132
+ represented the cyclical nature of time experienced from natural events such as weather patterns, rebirths, etc. The
133
+ third god Kairos represented the opportunist time, which reflects the appropriate time to achieve a task. In this way,
134
+ time, environment and beliefs were tightly linked with life and governed society and values.
135
+ On the other hand, in Eastern culture, the horoscope is one good example of a planetary and constellation frame-
136
+ work underpinning Astrology as a foundation of certain belief systems [20]. These beliefs assume that Astrology is
137
+ associated with time and causality, which can predict the future and guide humans.
138
+ The human observation of the night sky led to perceiving time from various cyclical patterns going far back in time.
139
+ For example, the Aboriginal Australians [21] observed the night sky and mapped them to the environment and life
140
+ stages that evolved various customs, arts and even religions. Not only by interacting planets and stars but the tilt of
141
+ the earth’s spin axis also significantly led to diversifying human cultures based on seasons, particularly when moving
142
+ away from the equator.
143
+ The notion of time and associated beliefs, customs, and values are important to consider when training AIS [22].
144
+ That will help promote human cultural values, ethics, and diversity, equity and inclusion (DEI). AI development may
145
+ need to pay attention to and integrate the time attributes that emerged from nature, values and cultures. Humans may
146
+ include them in the policies for rewarding self-learning AI algorithms (e.g., in MDP).
147
+
148
+ THE IMPLICATION OF BIOLOGICAL TIME ON AIS
149
+ The biological cycles play a fundamental role in human behaviours and the perception of time—for example, mood
150
+ cycles, circadian rhythms, and the menstrual cycle. Without understanding these biological time-keeping processes,
151
+ AI cannot seamlessly integrate with human society when creating values in health, culture, art, etc. These insights
152
+ are essential to realising emotional intelligence, empathy and awareness in AI. Literature shows the effective use of
153
+ Cyclic Hidden Markov Models (CyH-MMs) for detecting and modelling cycles in a multidimensional heterogeneous
154
+ biological time series data collection [23]. It is important to attribute the relevant features of biological processes
155
+ when training AIS, which raises more awareness about humans.
156
+ Recent discoveries in quantum physics argue that our reality is non-local, where awareness can happen instantly,
157
+ faster than the speed of light. Physicists and neurologists think brain neurons might be aware of the quantum world
158
+ through the orchestrated collapse of microtubules in the neurons in the brain [24, 25]. If this hypothesis is true, then
159
+ there are possibilities that human awareness can be linked with non-local realities to expand our consciousness across
160
+ the universe instantly. From this perspective, future AI might need to be evolved with the capabilities of biological
161
+ neurons, which interplay with the quantum realities. The recent development of neurotech realising brain-computer
162
+ interface (BCI) along with emerging quantum computers might enable such capabilities in the near future [26].
163
+ CONCLUSION
164
+ Consciousness and perception of time and causation are key to awareness and understanding reality. The notion of
165
+ time emerged from causation, a perception relative to the observer as per the relativity principles. In relativity, it’s
166
+ not time but the light-speed constant in all frames of reference. In contrast, in quantum entanglement, the reality is
167
+ non-local, and information can be transferred instantly faster than light. While the principles of time contradict the
168
+ foundation of physics, time also influenced the formation of diverse customs, values and cultures based on patterns
169
+ that emerged from nature, particularly around the regularities in the earth’s movement, environment, astronomy and
170
+ biology. Therefore, understanding time and related artefacts (i.e., cultures, beliefs, values, customs, physics, health,
171
+ etc.) are very important to realise deep awareness of reality. From the AIS perspective, it will enhance the understand-
172
+ ing of AI in human health, cultures, customs, values and various other diversities. Bringing this awareness to AI will
173
+ be a challenging and complex yet rewarding milestone in the evolution of ethical and responsible AI.
174
+ REFERENCES
175
+ 1. L. Kent and M. Wittmann, “Erratum to: Time consciousness: the missing link in theories of consciousness,” Neuroscience of Consciousness,
176
+ vol. 2021, 05 2021.
177
+ 2. M. Samarawickrama, “Unifying Matter, Energy and Consciousness,” International Conference on Mathematical Modeling in the Physical
178
+ Sciences (IC-MSQUARE), 9 2022. https://youtu.be/Pby1TfEluqE.
179
+ 3. A. Einstein, “Zur Elektrodynamik bewegter Körper. (German) [On the electrodynamics of moving bodies],” Annalen der Physik, vol. 322,
180
+ no. 10, pp. 891–921, 1905.
181
+ 4. S. Blackburn, “Hume and thick connexions,” Philosophy and Phenomenological Research, vol. 50, pp. 237–250, 1990.
182
+ 5. I. Newton, Philosophiae Naturalis Principia Mathematica. London: Royal Society, 1687.
183
+ 6. A. Einstein, “Zur allgemeinen Relativitätstheorie. (German) [Toward a General Theory of Relativity],” j-S-B-PREUSS-AKAD-WISS-2,
184
+ pp. 778–786, 799–801, 1915.
185
+ 7. A. Einstein, “Erklärung der Perihelbewegung des Merkur aus der allgemeinen Relativitätstheorie. (German) [Explanation of the perihelical
186
+ motion of Mercury from the General Theory of Relativity],” j-S-B-PREUSS-AKAD-WISS-2, pp. 831–839, 1915.
187
+ 8. J. A. Wheeler, “Geometrodynamics and the Problem of Motion,” Reviews of Modern Physics, vol. 33, pp. 63–78, Jan 1961.
188
+ 9. N. F. Ramsey, “History of early atomic clocks,” Metrologia, vol. 42, p. S1, jun 2005.
189
+ 10. R. Maddison and C. N. Mhurchu, “Global positioning system: a new opportunity in physical activity measurement,” International Journal of
190
+ Behavioral Nutrition and Physical Activity, vol. 6, no. 1, p. 73, 2009.
191
+ 11. P. A. Zyla et al. (Particle Data Group), “Review of Particle Physics,” Progress of Theoretical and Experimental Physics, vol. 2020, 08 2020.
192
+ 083C01.
193
+ 12. A. Einstein, B. Podolsky, and N. Rosen, “Can quantum-mechanical description of physical reality be considered complete?,” Phys. Rev.,
194
+ vol. 47, pp. 777–780, May 1935.
195
+ 13. A. Lohrey and B. Boreham, “The nonlocal universe,” Communicative & Integrative Biology, vol. 13, no. 1, pp. 147–159, 2020.
196
+ 14. W. Heisenberg, “Über den anschaulichen Inhalt der quantentheoretischen Kinematik und Mechanik. (German) [On the ideological content of
197
+ quantum theoretical kinematics and mechanics],” Z. Physik, vol. 43, pp. 172–198, Mar. 1927.
198
+ 15. D.
199
+ D.
200
+ Hoffman,
201
+ “The
202
+ Origin
203
+ of
204
+ Time
205
+ In
206
+ Conscious
207
+ Agents,”
208
+ Cosmology,
209
+ vol.
210
+ 18,
211
+ pp.
212
+ 494–520,
213
+ 2014.
214
+ https://www.cogsci.uci.edu/ ddhoff/HoffmanTime.pdf.
215
+
216
+ 16. W. Sulis, “Locality Is Dead! Long Live Locality!,” Frontiers in Physics, vol. 8, 2020.
217
+ 17. E. Oks, “Brief review of recent advances in understanding dark matter and dark energy,” New Astronomy Reviews, vol. 93, p. 101632, 2021.
218
+ 18. M. van Otterlo and M. Wiering, “Reinforcement Learning and Markov Decision Processes,” Reinforcement Learning: State-of-the-Art, pp. 3–
219
+ 42, 2012.
220
+ 19. J. E. Smith, “Time, Times, and the ‘Right Time’; Chronos and Kairos,” The Monist, vol. 53, no. 1, pp. 1–13, 1969.
221
+ 20. N. Campion, “Astrology as cultural astronomy,” Handbook of Archaeoastronomy and Ethnoastronomy, pp. 103–116, 2015.
222
+ 21. D. W. Hamacher, “Comet and meteorite traditions of aboriginal australians,” Encyclopaedia of the History of Science, Technology, and
223
+ Medicine in Non-Western Cultures, pp. 1–4, 2008.
224
+ 22. K. Lee and K. Joshi, “Understanding the Role of Cultural Context and User Interaction in Artificial Intelligence Based Systems,” Journal of
225
+ Global Information Technology Management, vol. 23, no. 3, pp. 171–175, 2020.
226
+ 23. E. Pierson, T. Althoff, and J. Leskovec, “Modeling Individual Cyclic Variation in Human Behavior,” in Proceedings of the 2018 World Wide
227
+ Web Conference, p. 107–116, 2018.
228
+ 24. S. Hameroff and R. Penrose, “Consciousness in the universe: A review of the ‘Orch OR’ theory,” Physics of Life Reviews, vol. 11, no. 1,
229
+ pp. 39–78, 2014.
230
+ 25. S. Hameroff, “‘Orch OR’ is the most complete, and most easily falsifiable theory of consciousness,” Cognitive Neuroscience, vol. 12, no. 2,
231
+ pp. 74–76, 2021.
232
+ 26. S. Saha, K. A. Mamun, K. Ahmed, R. Mostafa, G. R. Naik, S. Darvishi, A. H. Khandoker, and M. Baumert, “Progress in Brain Computer
233
+ Interface: Challenges and Opportunities,” Frontiers in Systems Neuroscience, vol. 15, Feb 2021.
234
+