Reproduce “Multi-Instance Causal Representation Learning for Instance Label Prediction and Out-of-Distribution Generalization” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Weijia Zhang, Xuanhui Zhang, Han-Wen Deng, Min-Ling Zhang
CausalMIL uses bag-level labels and shared bag context to learn instance representations for weakly supervised prediction. It conditions a latent prior on bag information and trains a classifier using one selected positive witness per bag. It achieves 0.878 ± 0.046 AUC-PR on Colon Cancer and 0.816 ± 0.015 accuracy on ColoredFashionMNIST OOD testing, under the reported experimental setups.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.