2024 ICML ICML 2024

Causally Motivated Personalized Federated Invariant Learning with Shortcut-Averse Information-Theoretic Regularization