video_id
stringlengths 11
11
| text
stringlengths 361
490
| start_second
int64 0
11.3k
| end_second
int64 18
11.3k
| url
stringlengths 48
52
| title
stringlengths 0
100
| thumbnail
stringlengths 0
52
|
|---|---|---|---|---|---|---|
oDKXwxaGkNA | bit no big modifications and your model is not going to give you the missing rectangle anymore but it's going to give you the age of the person that's visible on that image and all that by simply using a model that has acquired general understanding through self supervision and then you just fine tune it with little label data for the tasks you really want to solve what you gain from | 1,235 | 1,257 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1235s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | that is overall well first of all since your first building up general understanding and then fine-tuning your model to the task you want to solve and that means your model will need less label data to achieve better performance your model will converge faster because its weight already in a somewhat optimal position and also your model is going to be more robust because the features it | 1,257 | 1,279 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1257s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | learned these important characteristics will learn on a very large data set on yourself supervised data set and you were just fine-tuning on your downstream tasks where you're potentially a very little labeled data this is actually really a win-win situation here but how does this work in practice instead really usable in practice well and this is why I brought this industry case with | 1,279 | 1,301 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1279s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | me which we did together with upstream ability and our task at hand was estimating mode of transport from GPS movement data well GPS movement data is quite easily described you just have sequences of GPS coordinates and then associated timestamp and basically what we what we got was a data set of these trips across Vienna and that person here started his trip at the Green Dot ended it at the | 1,301 | 1,327 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1301s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | red dot and every dot in between was a GPS signal our task was for each of these GPS signals estimate which mode of transport he was taking was it going by bus was it going the Metro was he walking cycling going the car and so on actually quite hard quite complex given the little information so the first thing we asked our self here was so how would we humans solve this problem well what I | 1,327 | 1,352 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1327s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | would do is I would go on Google Maps and check out where transport lines are running right because if a GPS signal is nowhere close to a metro line probably he was not going by Metro and that's what we did we enriched this map of Vienna by adding image channels and each image channel represents a different transport network one represents the Metro network one the bus network and so | 1,352 | 1,373 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1352s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | on this is awful for humans right this produces an image you cannot interpret but it is great when it comes to deep learning well and then we used that information to embark on building a super lies deep learning model with no self supervision at this point the way we framed this problem was well we framed it as a sequence learning problem what does this mean so basically we | 1,373 | 1,396 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1373s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | thought our hypothesis was the movement history of a person provides us with important information also for estimating his current mode of transport because obviously if a person has been on the Metro for the last five GPS signals because his GPS signals were just on top of that metro line and his current trip ethical is again on top of that metro line it is quite likely that | 1,396 | 1,418 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1396s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | he is still on that Metro because changing transport modes of transport is actually associated with costs in terms of time effort and maybe even money so our hypothesis was abused the previous movement history to estimate this current mode of transport so again how do we frame this a machine learning problem we have a big big image with dots on it and then we want to | 1,418 | 1,441 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1418s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | incorporate the sequence of dots to make our estimation so you can do this in a variety of ways of course the way we did it was for each of these dots we want to incorporate in our for making our prediction we cropped out a small rectangle of the image around it so we end up with a sequence of image tiles each of these image tiles representing the path in total then goes into a | 1,441 | 1,465 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1441s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | convolutional neural network a so-called CNN which is a type of neural network that is very good at extracting important information from images that CNN then produces what that CNN thinks is important for making the modality estimation so called feature vectors it learns feature vectors which is simply a numeric dense vector representation of that image and what is important in that | 1,465 | 1,488 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1465s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | small image title in that sequence of image a feature vector s then goes into a recurrent neural network a so called R and n an errand n again is a type of neural network just a very different type which is great for modeling sequences from modeling temporal dependencies between elements of a sequence and since we are dealing with a sequence of GPS coordinates here or a | 1,488 | 1,510 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1488s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | sequence of image tiles are actually now a sequence of learned image features an RNN is an obvious choice here and there are an N then finally it gives you in the end which modality the person was likely to be taking was he walking going by bus taking the metro and so on and that what you can see here all that was basically one single deep learning model and it worked out reasonably well | 1,510 | 1,535 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1510s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | actually surprisingly well however you can imagine that this was not a small model actually we were dealing with I think more a bit more than hundreds of thousands of parameters in that model which is ok if you have sufficient labeled data and you can guess well we didn't of course we were dealing with rather little real-world data here which is a problem given if you use a | 1,535 | 1,558 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1535s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | large network because the network is likely to overfit and so on so we thought about well how about self supervised deployer this is actually a very good very good case for that how about first letting the model develop general understanding and then taking this model with all of its understanding and using it for solving our very specific modality estimation tasks and | 1,558 | 1,580 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1558s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | that's how we came up with that concept you could for example teach your model the entire Vienna transport network in a self supervised pre training stage just as we did before then take the not take that model with all of its knowledge and find unit to make modality estimation happen sounds good but how do you teach a model EBN a transport network and this is again | 1,580 | 1,602 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1580s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | where it comes to self supervision where it comes to creativity how can you frame a problem from the data you have to make this happen one way is to for example randomly sample 32 by 32 pixels again these small image tiles from a map of Vienna which is publicly available of course and then you just automatically check which line of transport is running on that tile | 1,602 | 1,627 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1602s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | it's all information is publicly available again you take that as labels and you teach a model to predict which line of transport is running on a given image tile your model is going to learn the Vienna transport network it's going to be an absolute expert and you can then take this model you can take this CNN that knows everything about the Vienna transport network and put it back | 1,627 | 1,648 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1627s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | into our original architecture replacing the CNN from before the CNN that had to be trained from scratch our model now does need to be trained from scratch it has a big big head start it has a quiet general understanding so it just needs to be fine-tuned a bit and that's an end and passes on high-quality feature vectors again to RNN which learns the temporal dependencies and outputs the | 1,648 | 1,671 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1648s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | predictions what you gain from this is you have a significantly smaller model because you do not need to train the CNN from scratch this means you can have many of the parameters frozen and you end up with a more with a model with trainable parameters going from a hundred thousand before to ten or tens of thousands now which is a lot better suited for little labelled data so again | 1,671 | 1,695 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1671s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | what you're going to get is your model will be able to deal with little label data better it's gonna be more robust it's gonna converge faster because you simply have to optimize less parameters here and overall it's gonna make more accurate predictions given the same labelled data we have come a pretty long way now all the way from supervised deep learning why it's so great so powerful | 1,695 | 1,716 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1695s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | but also we got to know its ugly side we got to know that it just did tons of label data to achieve this famous performance subsequently we found out how taking a self supervised approach can help us overcome that key problem of supervised learning and finally we're looking through the industry case that hopefully gave you a pretty good practical idea of how you can apply self | 1,716 | 1,737 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1716s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | supervision to your problems to make your models work better on little label data I hope you enjoyed the talk feel free to ask any question you might have in case we are out of time for questions I will be hereafter talk as well so feel free to approach me and ask any question in private that may have been left unanswered also of course drop me an email and add me on LinkedIn | 1,737 | 1,759 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1737s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | I'm always happy for for method for interesting message as I get and most of all visit our website craft works dot AI thank you very much okay um let's see I think we maybe have time for one question would you like to choose which one you think would be the best one the unites answer okay I think the second one what is the main difference between supervisors and self | 1,759 | 1,791 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1759s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | supervised learning this is a important question because it's very fundamental so self supervised learning is supervised learning the concept behind self supervision is that you do not have a label dataset but you come up with a smart task how you can create labels from a previously unlabeled data set not needing any human labeling this is the concept of self supervision you use the | 1,791 | 1,817 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1791s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
oDKXwxaGkNA | data that you have that video that we had before you do not have any labels you cannot just apply supervised learning because you don't have labels but you need to implement a self supervised task that where you first pull out these labels to actually make supervised learning happen and you're learning from a distribution that is within the data actually thank you and | 1,817 | 1,839 | https://www.youtube.com/watch?v=oDKXwxaGkNA&t=1817s | Self-Supervised Learning - Towards Autonomously Learning Machines—Simon Stiebellehner | |
LHXXI4-IEns | hi and welcome to an illustrated guide to recurrent neural networks a Michael also known as learned vector I'm a machine learning engineering the natural language processing and voice assistance space if you're just getting started in machine learning and want to get some intuition behind recurrent neural networks those videos for you if you want to get into machine learning | 0 | 18 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=0s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | recurrent neural networks are a powerful technique that's important to understand if you use smart phones and frequently surf the internet odds are you use applications that leverages RN ends recurrent neural networks are using speech recognition language translation stock prediction it's even using image recognition to describe it content in pictures so I know there are many guys | 18 | 42 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=18s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | on recurrent neural networks but I want to share illustrations along with an explanation of how I came to understand it in this video I'm going to avoid all the math and focus on the intuition behind RN ends instead by the end of this video you should have a good understanding of RN ends and hopefully have that light bulb moment so RN ends are neural networks that are good at | 42 | 63 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=42s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | modeling sequence data to understand what that means let's do a thought experiment say you take a still snapshot of a ball moving in time let's also say you want to predict a direction that the ball is moving so with only the information that you see on the screen how would you do this well you can go ahead and take a guess but any answer you come up with would be that a random | 63 | 86 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=63s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | guess without knowledge of where the ball has been you weren't having an update of to predict where it's going if you record many snapshots of the balls position in succession you will have enough information to make a better prediction so this is a sequence a particular order in which one thing follows another with this information you can now see that the ball is moving | 86 | 107 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=86s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | to the right sequence data comes in many forms audio is the natural sequence you can chop up an audio spectrogram into chunks and feed that into RN ends text is another form of sequences you can break text up into sequence of characters or sequence of words okay so our ends are good at processing sequence data for predictions but how well they do that by having a concept I | 107 | 134 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=107s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | like to call sequential memory to get a good intuition behind what sequential memory means I like to invite you to say the alphabet in your head go on give it a try that was pretty easy right if you were taught the specific sequence it should come easily to you now try saying the alphabet backward I bet that was much harder unless you practice the sequence before you'll likely have a | 134 | 162 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=134s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | hard time here's a fun one start out the letter F at first just struggle with the first few letters but then after your brain picks up the pattern the rest will come naturally so there's a very logical reason why this can be difficult you learn the alphabet as a sequence sequential memory is a mechanism that makes it easier for your brain to recognize sequence patterns all right so | 162 | 186 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=162s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | are n ends have this abstract concept of sequential memory but how the heck does it replicate that concept well let's look at a traditional neural network also known as a feed-forward neural network as an input layer hidden layer and output layer how do we get a feed-forward neural network to be able to use previous information to affect later ones what have we had a loop in a neural | 186 | 207 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=186s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | network that can pass previous information forward and that's essentially what a recurrent neural network does an RNN has a looping mechanism that acts as a highway to allow information to flow from one step to the next this information is the hidden state which is a representation of previous inputs let's run through an art and use case to have a better understanding of how this works let's | 207 | 230 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=207s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | say we want to build a chatbot they're pretty popular nowadays let's say the chatbox can classify intentions from the user's inputted text to tackle this problem first we're going to encode the sequence of texts using an RNN then we're going to feed the RNA and output into a feed-forward neural network which will classify the intents okay so a user types in what time is it | 230 | 254 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=230s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | to start we break up the sentence into individual words rnns work sequentially so we feed it one word at a time the first step is to feed what into the RNA the RNA encode what and produces an output for the next time we beat the work time in a hidden state from the previous step remember that the hidden state represent information from all previous steps the RNN now has | 254 | 279 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=254s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | information about the work what in time we repeat this process until the final step you can see about a final step the Arnon has encoded information from all the words in the previous steps since the final output was created from the rest of the sequence we should be able to take the final output and pass it to the feed-forward layer to classify in intent for those of you who like looking | 279 | 303 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=279s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | at code here are some Python showcasing the control flow first you initialize your network layers in the initial hidden state the shape and dimensions of the hidden state will be dependent on the shape and dimension of your recurrent rail network then you loop through your inputs past a word and hence a into the artnet DRN and returns the output at a modified hidden state | 303 | 325 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=303s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | this modified hidden state should now contain information from all your previous steps you continue to loop until you're out of words last you pass the output to the feed board layer and it returns a prediction and that's it the control flow of doing a forward pass of a recurrent neural network is a for loop okay now back to our visualization you may have noticed the odd | 325 | 349 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=325s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | distribution of colors in the hidden states this is to illustrate an issue with our n ends known as short-term memory short-term memory is caused by the infamous vanishing gradient problem which is also prevalent in other neural network architectures so as yarn and processes more steps it has troubles retaining information from previous steps as you can see the | 349 | 369 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=349s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | information from the word what and time is almost non-existent at the final step short-term memory and vanishing gradient is due to the nature of back propagation algorithm used to Train and optimize neural networks to understand why this is let's take a look at the effects of back propagation on a deep feet board neural network training and neural network has three major steps first it | 369 | 392 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=369s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | does a forward pass and makes a prediction second it compares the prediction to the ground truth using a loss function the loss function outputs an error value which is an estimate of how badly the network is performing last it uses the error value to do back propagation which calculates the gradients for each node in the network the gradient is a value used to adjust | 392 | 414 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=392s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | the network's internal weights allowing the network to learn the bigger the gradient the bigger the adjustments and vice versa here's where the problem lies when doing back propagation each node in a layer calculates its gradient with respect to the effects of the gradients and the layer before so the adjustments in the layer before it is small then the adjustments in the current layer will be | 414 | 435 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=414s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | even smaller this cost gradients to exponentially shrink as it back propagates down the earlier layers failed to do any learning as the internal weights are barely being adjusted due to extremely small gradient and that's the vanishing gradient problem let's see how this applies to recurrent neural networks you can think of each time step and over current no | 435 | 457 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=435s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | network as a layer to train a recurrent neural network use an application of backpropagation called back propagation through time the gradients value will exponentially shrink as it propagates for each time step again the gradient is used to make the adjustments in the neural networks weights thus allowing it to learn small gradients means small adjustments this causes the early layers | 457 | 480 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=457s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | to not learn because of the vanish ingredients the RNN doesn't learn the long-range dependencies across time steps this means that there is a possibility that the word light and time are not considered when trying to predict a user's intention the network has to make its best guess with is it that's pretty ambiguous and would be difficult even for a human so not being | 480 | 504 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=480s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | able to learn on earlier time steps causes the network tap short-term memory okay so RNN suffer from short-term memory so how do we combat that to mitigate short-term memory to specialized recurrent neural networks were created one called long short term memory or lsdm for sure the other is gated recurrent units or gr use LS TMS and gr use essentially fokin just like our meds but they're capable | 504 | 531 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=504s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | of learning long-term dependencies using mechanism called gates these gates are different tensor operations that can learn what information to add or remove to the hidden state because of this ability short-term memory is less of an issue for them to sum this up RNs are good for processing sequence data for predictions but suffer from short-term memory the short-term memory issue for | 531 | 553 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=531s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
LHXXI4-IEns | vanilla arm ends doesn't mean to skip them completely and you still more involved versions like LST MS or gr use RNs have the benefit of training faster and uses less computational resources that's because there are less tensor operations to compute you could use LST MS or gr use when you expect a model longer sequences with long-term dependencies if you're interested in | 553 | 574 | https://www.youtube.com/watch?v=LHXXI4-IEns&t=553s | Illustrated Guide to Recurrent Neural Networks: Understanding the Intuition | |
NXaVLXSZdEw | Germany's super-rich no other European country has as many billionaires and while their fortunes are growing more and more Germans are living under the poverty line set by the Organisation for Economic Cooperation and Development the OECD the press frequently reports on the country's high income inequality and low social mobility but little is known about the super-rich nate'd sure money | 0 | 26 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=0s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | attracts success and success attracts money I do believe that I've experienced myself you can suddenly connect with people previously out of reach who are the kind of shores that had - who are Germany's super wealthy how do they live and how do they see the country they live in the way wealthy people in Germany are talked about bothers me because it's sensationalizing in sycophantic and in | 26 | 55 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=26s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | no way reflects what wealthy people have done and continue to do for this country it creates this impression of rich people being like Scrooge McDuck that they had these money bins in which they wallow in their coins look Scrooge McDuck wants more more MORE that's not my world I like earning money is less but not swimming in it we thought we wanted to get closer to the discrete | 55 | 85 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=55s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | world of Germany's ultra rich of company owners and heirs worth millions we wanted to find out what makes those on top of the world tick [Music] every year an exclusive event takes place in the Schloss hotel Kornberg near frankfurt to which the public is not invited it's the annual Hall of Fame evening for the business monthly manager magazine hardly any other occasion in | 85 | 126 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=85s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | Germany draws as many wealthy business owners [Music] the Iggy had laudatory esteemed presenters dear jury members dear ladies and gentlemen welcome to manager magazine's Hall of faith when we first founded the Hall of Fame in 1992 we wanted to take a stand for excellence and unconditional entrepreneurship and against faintheartedness and averageness we begin our nominations today with Harv | 126 | 159 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=126s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | dama moot his company United Internet is valued at around 11 billion and on our list of Germany's richest people he ranks 25th with a personal wealth of 4.5 billion euros collectively the guests of this elegant evening are worth billions of euros this is the face of wealth in Germany mainly male and although we're close to the super-rich here their world remains | 159 | 190 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=159s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | somehow out of reach Hamburg home to manager magazine's parent company the speaker group hockey any publication keeps closer tabs on Germany's ultra rich every year the magazine's team gathers information on wealthy Germans and using the Forbes model makes a special edition with a list of Germany's 1001 richest people it's painstaking work how many billionaires are video | 190 | 231 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=190s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | last year it was a hundred and thirty six right no last year it was around 170 170 okay someone who wants to get on our list at 1001 richest Germans needs to have around 100 million that doesn't have to be money in the bank most people have it as assets or as property but that's the ballpark we're looking for to be on our list of the richest Germans the thousand vison contant | 231 | 256 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=231s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | so nice and our chances editor-in-chief Stephon clothespin has been around Germany's ultra-wealthy for years what does it take to get on the list as it Amazon the top I'd say that the top 150 spots on our list will always go to company owners and heirs even if you're a chief physician you'll need to see a whole lot of patients to become a billionaire managers also have | 256 | 286 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=256s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | a hard time getting that high up here in Germany there's a debate about whether the heads of Dax companies earn too much but if you compare it to what people in similar posts in the US the UK or China earn its peanuts so that alone can't ever make you one of the truly ultra rich in them it's a club des vins a new chef creating a list of the wealthiest people | 286 | 315 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=286s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | is especially difficult in Germany though not because they aren't enough of them [Music] he somehow managed to amass this huge Empire and fortune in just a few years his wealth is estimated to be around four billion he's the least known super-rich Germans would you let us interview him at home they prefer to fly under the radar to avoid envy the super-rich don't like to melt themselves | 315 | 346 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=315s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | so to speak you can normally only get interviews inside the homes it's second-tier once there you'll always get guys who get a kick out of publicity like mr. Marsh - and such getting a good shot of mr. Marsh Mayer is never a problem but the real money keeps itself hidden the super-rich are trying to go unnoticed sometimes they even try to hide there are no photos of | 346 | 380 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=346s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | several people on our list you won't find a single picture if you go online and google them there hasn't been a photo of their hymens germany's richest family for decades many stay hidden because they want to live normal lives and think that they won't be able to do so if they're known to be multimillionaires Lemkin for months our interview requests were rejected agreed film shoots were | 380 | 405 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=380s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | canceled last minute none of the rich wanted to talk to us about money finally we got lucky in the financial hub Frankfurt here in a prime location tucked behind the Bank towers sits the asset management company for come Foca managers German business family's fortunes its chairman Christian fanbase Todd's home provided some insight as to why rich Germans are so shy | 405 | 433 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=405s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | the joy many wealthy Germans are reluctant about stepping out into the public eye because they're afraid they could be seen negatively they ask themselves what do I get from showing myself to the public it doesn't give me anything on the contrary it could lead to some crazy person taking note and breaking into my home or kidnapping one of my children and those fears are not | 433 | 458 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=433s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | unwarranted then there's also the fact is that many heirs are inheriting fortunes that are somehow tainted by or related to the Third Reich the Abbott ear of rich Germans avoids publicity as if it were the plague but what are they so afraid of if I asked a number of my friends whether they'd like to be interviewed for this film and each one said no they'd say someone else can do | 458 | 496 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=458s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | that better than me I can't do it right and I might come across wrong they think they'd have much more to lose than to gain even calm [Music] after a lot of back and forth with his press team one ultra-rich German did agree to meet us Michele Otto is the chairman of the Supervisory Board of the Auto Group and one of the ten richest Germans we asked him why Germans are reluctant | 496 | 526 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=496s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | to show off their wealth any wiring doing so would lead to envy in the u.s. achievement and wealth has seen in a highly positive way but here they carry a bitter aftertaste where does his fortune come from or how did he get his wealth that scares some artists whom with a shrine weren't you I think a lot of people find their own wealth a little nauseating I like it Pindos good dick | 526 | 559 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=526s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | Hoffman grew up poor his mother ran a small drugstore in post-war Hanover her son had bigger plans in 1972 the idea of opening the first self-service drugstore in Germany came to him today he's a multi-billionaire why was he happy to step into the spotlight negan unfunctional in the early years it was just about getting the name rossmann out there so when I was invited to a talk show on a small | 559 | 589 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=559s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | regional channel I liked going because I thought free publicity for my company but then two or three years ago I started to understand that this slightly flabby balding man whose teeth aren't perfect was hungry for recognition himself back then I thought I was stepping into the limelight to promote the company but everyone rationalizes their motives and I did too y'all stop | 589 | 618 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=589s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | adore within 40 years horsemen became the most profitable drug store chain in Europe with stores in six countries in 2018 a total of 56 thousand people were working on the chain Hoffman is active in other business areas as well and he speculates on the stock market that could not take any capitai legal debt so I have a couple of private equity investments the | 618 | 658 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=618s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | largest is palliated between 80 and 100 million depending on the spot rate so it's a fair sum there are also shares in different industries introduced a mob will be boring it's off to tear a little I'm fine you're still referred to as an SME because small medium enterprise right why do you think that so I don't really know with 55,000 employees you're not really a medium enterprise anymore | 658 | 687 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=658s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | you're in another league we're back at Manager magazine in Hamburg with photos with a special issue are being selected the list of the richest Germans includes a notably high number of company owners from the so called medium-sized businesses it's a German peculiarity and it's not the only one the cover story for this current issue is that last year money rained | 687 | 716 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=687s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | down on Germany's super-rich up here we have the chef Lutz the matriarch and her son who now owns 80% of mom has kept only 20 on the chef's loose have been at the top of our list for many many years I'd have to check exactly how much they're worth but around 20 billion give or take a bit down here we have MS Basel toss in the head of the Hinkle class she's the first and so far only woman | 716 | 742 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=716s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | chairman of the Supervisory Board of one of the 30 Dax companies Germany's economy is still extremely male-dominated what's interesting is if you compare our list here with a list of the ultra-rich in the US we have a lot of old money old companies that have been around for decades in the US you have all those lads from Facebook Google snapchat and so forth that have bubbled | 742 | 766 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=742s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | up to the top of the list we don't have that type of thing here and compared with other countries Germans are very reticent about showing their wealth very few Germans sail around in boats like this usually that's Americans Russians Chinese and so on here you don't really show your money you might have various houses villas and such but there's likely to be a Fulks park and | 766 | 790 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=766s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | parked out front it seems the average ultra wealthy German Israeli conspicuous unlike in the US athletes actors and TV personalities rarely make it onto the German list important on guns even though we put a great deal of love in a sweat into estimating guests fortunes they're probably much bigger especially if we're talking about urban real estate prices have exploded over the past 10 years | 790 | 819 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=790s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | a lot of people have doubled their property assets so if you started with five billion in that market you're likely to have ten or fifteen billion today money makes money but while rich Germans fortunes have exploded since the financial crisis due to the increase in value of real estate stocks and assets those with average incomes have had to swallow losses even liberal economics | 819 | 846 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=819s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | Institute's are concerned about social inequality in Germany how do the super-rich see this disparity the headline in the newspaper reads the rich are getting richer and richer which is true the rich are getting richer yes it's true but it's also false there are 20 million citizens in Germany who have assets worth between 100,000 and 1 million so millions of people are getting richer | 846 | 875 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=846s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | now the rich are getting richer even faster because one factor is probability that they can dedicate much more time to increasing their wealth about initiative the Left Party would likely say split up your wealth but my response is I also do things for the world which I live in I don't just take I'm not a socialist though I can only do things because I have things I can y'know Bastogne Posada | 875 | 904 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=875s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | Michel Otto is one of the rich whose wealth has increased he successfully transitioned his mail-order business into a digital enterprise over 20 companies now belong to the Auto Group we wanted to know how Otto sees the debates about rich and poor do the rich understand the worries of the poor in the tough one just one good when people talk about those on top who don't | 904 | 930 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=904s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | understand those at the bottom I wouldn't say that applies to me because I was not born rich I came to Hamburg as a refugee from West Prussia and my father had to start from scratch that's why I absolutely do understand people living in poverty today on the other hand I'd also say that if Germany is getting more and more millionaires people with small or medium-sized | 930 | 957 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=930s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | businesses because generally the millionaires in question did build a business they now run then I think that's great because they are the people creating jobs for me that's what should matter in this debate we should focus on that and not rich versus poor voted is good young ensuring that the wealth of rich families can increase despite current zero interest rates is the mission of | 957 | 981 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=957s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | Christian fund based on science company each day he and his employees sent out investment opportunities from banks and other entities given exclusively to his wealthy clients this is a way to present an offer to just a handful of valued clients and focus Oh [Music] is it fair that rich Germans are able to increase their wealth while the rest of the population gets left behind | 981 | 1,022 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=981s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | [Music] I think it's difficult to apply terms like equity and fairness to the distribution of wealth I would say that here in Germany were better off than ever before and people here live in above-average circumstances nonetheless we have to make sure the gap between rich and poor doesn't get too wide because we don't want social conflicts like in the US or Latin America to | 1,022 | 1,048 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1022s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | happen here for better times company Fulgham is a so-called multi-family office family officers take care of the needs of very wealthy families managing and increasing their assets it only makes sense to use family offices if your wealth is upward of 30 million euros who can afford such a thing well a family office like ours obviously can't discuss its clients we | 1,048 | 1,077 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1048s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | have well known German business families that's our typical client profile Mikasa she can take someone who is or was a company owner thinks differently to someone who has spent their life as an employee or they are families who have had money for a long time so wealth is so to speak in their genes and then there are families who have just come into their wealth people who are still | 1,077 | 1,105 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1077s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | pumping with entrepreneurial energy they're usually quite different from heirs option Kaka I see it as a top top in my Stanford [Music] - Allah seems to have plenty of this entrepreneurial energy his company is McFate Europe's largest fitness studio chain its headquarters are in an old baking Factory in Berlin Hina Shula started off small but today his wealth | 1,105 | 1,134 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1105s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | is valued at 250 million euros [Music] which bring me finish then they're from I was born near long back grew up in a small village normally in villages you do the sports available and when I was around 1516 my role models were Arnold Schwarzenegger and Stallone that's how I came to the fitness world Shahla went from secondary school to complete a Salesman apprenticeship and | 1,134 | 1,164 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1134s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | became the manager of three supermarkets then he decided to start something new - when I was 25 I decided to switch to the fitness industry my idea was to open a gym where anyone could train the matter kids or her income that was the initial idea and I had big goals I wanted to be number one in Europe but that's all I had I didn't have financing okay good it was 1997 when shala opened his first | 1,164 | 1,193 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1164s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | fitness studio in verts book close to his home village he made use of some unusual methods on the monetary hoods book was a big step for me I opened my first fitness studio there under the slogan now also inverts book which people saw through as a marketing gag because customers came to me and asked where else we had studios so I gave it some thought and came up with the next marketing gag soon | 1,193 | 1,222 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1193s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | also in ela anything that did the job and also put me on track to going from verse bog to Erlangen and then it grew from there cuza ten years later Heine Shula reached his goal he's number one in Europe and still expanding McFate now owns ten Fitness companies as well as its own model agency meanwhile Shanna is getting ready to open fitness studios in the US | 1,222 | 1,251 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1222s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | des moines severan is getting to be number one is much easier than staying number one I think if you want to be successful you need to be a bit of an alpha animal inside the investor would probably pick a brand and say okay that could work I like it although it's probably two steps too far from any but if someone doesn't want to get involved with us because of it so what I'm | 1,251 | 1,284 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1251s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | convinced that you have an easier time if you fought your way to the top and to success lions feel it and so do partners I think that's our situation which is why I can also imagine that someone who inherits something or takes over or even has to take over a company in the second or third generation will have a much tougher time and Montezuma Muhsin as CSV aha but there are plenty of heirs in | 1,284 | 1,312 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1284s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | Germany huge fortunes and thousands of companies have been passed from one generation to the next they are heirs who don't want to and others who shouldn't take over their parents businesses succession is hugely important among Germany's richest Michel Otto inherited the mail-order company from his father and successfully managed it from when the kinder is this it will | 1,312 | 1,336 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1312s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | be harder for my children though because now the Auto Group has a hundred and twenty three companies in over 30 countries I know every single company either because I was involved when it was founded or because I led the takeover negotiation but my children don't yet know that many companies guns wunderkind otto's children have opted against direct succession I think it's important | 1,336 | 1,364 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1336s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | to give your children the option without pressuring them so you don't force them into a role I think that mistake is made often and I'd say it's bad for both the children and the company when master kundo under Thunder Lim kaput so now we have one point of sale Berg Vader are you a bit tired yes it's for an exhausting couple of days in ten minutes I'll be fine again it was just Bam Bam | 1,364 | 1,390 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1364s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | Bam Bam dear horsemen also spent a lot of time considering who his successor would be it's now decided how old horsemen the younger of his two sons will take over as manager of the drugstore chain do we actually sell much yarn it's not exactly part of a drugstores range the trend sort of over it was big in 2016 already fading in 2017 and it's been stagnating in 2018 | 1,390 | 1,418 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1390s | Germany: The discreet lives of the super rich | DW Documentary | |
NXaVLXSZdEw | but it's not the worst product we have and it still brings in some revenue rose father had to show him the appeal of being in charge when the boys got more engaged I thought oh now we can't look as if the drugstore business is only about making money so I showed them how we're also active in social issues in Africa and so forth I've always showed my sons that what we do isn't just about | 1,418 | 1,445 | https://www.youtube.com/watch?v=NXaVLXSZdEw&t=1418s | Germany: The discreet lives of the super rich | DW Documentary |