DecAug: Out-of-Distribution Generalization via Decomposed Feature Representation and Semantic Augmentation

TL;DR

DecAug introduces a novel approach for out-of-distribution generalization by disentangling category and context features and applying semantic augmentation to enhance robustness across diverse distribution shifts.

Contribution

It proposes a decomposed feature representation method and gradient-based semantic augmentation to improve OoD generalization across various types of distribution shifts.

Findings

DecAug outperforms state-of-the-art methods on multiple OoD datasets.

Disentangling features improves robustness to distribution shifts.

Gradient-based augmentation enhances model generalization.

Abstract

While deep learning demonstrates its strong ability to handle independent and identically distributed (IID) data, it often suffers from out-of-distribution (OoD) generalization, where the test data come from another distribution (w.r.t. the training one). Designing a general OoD generalization framework to a wide range of applications is challenging, mainly due to possible correlation shift and diversity shift in the real world. Most of the previous approaches can only solve one specific distribution shift, such as shift across domains or the extrapolation of correlation. To address that, we propose DecAug, a novel decomposed feature representation and semantic augmentation approach for OoD generalization. DecAug disentangles the category-related and context-related features. Category-related features contain causal information of the target object, while context-related features describe the attributes, styles, backgrounds, or scenes, causing distribution shifts between training and test data. The decomposition is achieved by orthogonalizing the two gradients (w.r.t. intermediate features) of losses for predicting category and context labels. Furthermore, we perform gradient-based augmentation on context-related features to improve the robustness of the learned representations. Experimental results show that DecAug outperforms other state-of-the-art methods on various OoD datasets, which is among the very few methods that can deal with different types of OoD generalization challenges.