Consistency Regularization for Domain Generalization with Logit Attribution Matching

Han Gao, Kaican Li, Weiyan Xie, Zhi LIN, Yongxiang Huang, Luning Wang, Caleb Chen Cao, Nevin L. Zhang

Keywords: domain generalization, consistency regularization, causality

TL;DR: We introduce Logit Attribution Matching, a new domain generalization technique using semantic sharing pairs that consistently outperforms existing methods

Abstract: Domain generalization (DG) is about training models that generalize well under domain shift. Previous research on DG has been conducted mostly in single-source or multi-source settings. In this paper, we consider a third lesser-known setting where a training domain is endowed with a collection of pairs of examples that share the same semantic information. Such semantic sharing (SS) pairs can be created via data augmentation and then utilized for consistency regularization (CR). We present a theory showing CR is conducive to DG and propose a novel CR method called Logit Attribution Matching (LAM). We conduct experiments on five DG benchmarks and four pretrained models with SS pairs created by both generic and targeted data augmentation methods. LAM outperforms representative single/multi-source DG methods and various CR methods that leverage SS pairs. The code and data of this project are available at https://github.com/Gaohan123/LAM.

List Of Authors: Gao, Han and Li, Kaican and Xie, Weiyan and Lin, Zhi and Huang, Yongxiang and Wang, Luning and Cao, Caleb and Zhang, Nevin

Code Url: https://github.com/Gaohan123/LAM

Submission Number: 15