Repeated Environment Inference for Invariant Learning
This work addresses a key challenge in invariant learning for machine learning practitioners by providing a method to infer environments when labels are unknown, though it is incremental as it builds on existing environment inference frameworks.
The paper tackles the problem of invariant learning without known environment labels by proposing a repeated environment inference process that iteratively refines a reference model to better capture spurious correlations, resulting in improved invariant learning performance on synthetic and real-world datasets.
We study the problem of invariant learning when the environment labels are unknown. We focus on the invariant representation notion when the Bayes optimal conditional label distribution is the same across different environments. Previous work conducts Environment Inference (EI) by maximizing the penalty term from Invariant Risk Minimization (IRM) framework. The EI step uses a reference model which focuses on spurious correlations to efficiently reach a good environment partition. However, it is not clear how to find such a reference model. In this work, we propose to repeat the EI process and retrain an ERM model on the \textit{majority} environment inferred by the previous EI step. Under mild assumptions, we find that this iterative process helps learn a representation capturing the spurious correlation better than the single step. This results in better Environment Inference and better Invariant Learning. We show that this method outperforms baselines on both synthetic and real-world datasets.