Information Maximization for Long-Tailed Semi-Supervised Domain Generalization

TL;DR

This paper introduces IMaX, a novel method based on the InfoMax principle, designed to improve long-tailed semi-supervised domain generalization by effectively handling class imbalance and enhancing existing SSDG techniques.

Contribution

IMaX is a new objective that maximizes mutual information with an {\alpha}-entropic term, improving SSDG performance on long-tailed class distributions.

Findings

IMaX consistently improves performance of state-of-the-art SSDG methods.

The {\alpha}-entropic objective better handles class imbalance.

Empirical validation across two image modalities confirms effectiveness.

Abstract

Semi-supervised domain generalization (SSDG) has recently emerged as an appealing alternative to tackle domain generalization when labeled data is scarce but unlabeled samples across domains are abundant. In this work, we identify an important limitation that hampers the deployment of state-of-the-art methods on more challenging but practical scenarios. In particular, state-of-the-art SSDG severely suffers in the presence of long-tailed class distributions, an arguably common situation in real-world settings. To alleviate this limitation, we propose IMaX, a simple yet effective objective based on the well-known InfoMax principle adapted to the SSDG scenario, where the Mutual Information (MI) between the learned features and latent labels is maximized, constrained by the supervision from the labeled samples. Our formulation integrates an {\alpha}-entropic objective, which mitigates the class-balance bias encoded in the standard marginal entropy term of the MI, thereby better handling arbitrary class distributions. IMaX can be seamlessly plugged into recent state-of-the-art SSDG, consistently enhancing their performance, as demonstrated empirically across two different image modalities.