Adversarial Invariant Feature Learning with Accuracy Constraint for Domain Generalization

Abstract

Learning domain-invariant representation is a dominant approach for domain generalization, where we need to build a classifier that is robust toward domain shifts. However, previous methods based on domain invariance overlooked the underlying dependency of classes on domains, which is responsible for the trade-off between classification accuracy and domain invariance. This study proposes a novel method adversarial feature learning under accuracy constraint (AFLAC), which maximizes domain invariance within a range that does not interfere with classification accuracy. The reason for the constraint is that the primary purpose of domain generalization is to classify unseen domains rather than the invariance itself, and improving the invariance can negatively affect that performance. Empirical validations show that the performance of AFLAC is superior to that of baseline methods, supporting the importance of considering the dependency and the efficacy of the proposed method to overcome the problem.