Equivariant Disentangled Transformation for Domain Generalization under Combination Shift

TL;DR

This paper introduces an algebraic framework and a novel data augmentation method called equivariant disentangled transformation (EDT) to improve domain generalization when encountering unseen domain-label combinations, outperforming invariance-based approaches.

Contribution

It provides a formal algebraic formulation of combination shift and proposes EDT, a method leveraging equivariance and disentanglement for better domain generalization.

Findings

EDT improves performance over invariance-based methods.

Exploiting equivariance structures is crucial for handling combination shift.

Algebraic formulation clarifies the limitations of invariance in domain generalization.

Abstract

Machine learning systems may encounter unexpected problems when the data distribution changes in the deployment environment. A major reason is that certain combinations of domains and labels are not observed during training but appear in the test environment. Although various invariance-based algorithms can be applied, we find that the performance gain is often marginal. To formally analyze this issue, we provide a unique algebraic formulation of the combination shift problem based on the concepts of homomorphism, equivariance, and a refined definition of disentanglement. The algebraic requirements naturally derive a simple yet effective method, referred to as equivariant disentangled transformation (EDT), which augments the data based on the algebraic structures of labels and makes the transformation satisfy the equivariance and disentanglement requirements. Experimental results demonstrate that invariance may be insufficient, and it is important to exploit the equivariance structure in the combination shift problem.