Questions about Agata aggregator
Dear Community, I am interested in the details about the Agata aggregator, but was not able to find enough information about it, e.g. the architecture description, the detailed layout design of the layers of the Aggregator, etc. Could you please advise? I also dig into the reference (Raciti, P. et al. Clinical validation of artificial intelligence–augmented pathology diagnosis demonstrates significant gains in diagnostic accuracy in prostate cancer detection. Arch. Path. Lab. Med. 147, 1178–1185 (2022)) noted in the the Virchow article in 'Nature', but didn't find detailed introduction about Agata in the reference either. Any help would be highly appreciated!
Also, it seems the Extended Data Fig. 1 and 2 are mismatching with the captions in the Nature article. Thanks.
Thanks for your interest. You can find more about the Agata aggregator in our aggregator training library here: https://github.com/Paige-AI/paige-ml-sdk/
Specifically, the model definition can be found in this file: https://github.com/Paige-AI/paige-ml-sdk/blob/main/src/paige/ml_sdk/model_universe/nn/agata.py
Hope this helps!
Great, this is very helpful, thanks, much appreciated!