SelfReg: Self-supervised Contrastive Regularization for Domain Generalization

TL;DR

SelfReg introduces a self-supervised contrastive regularization technique for domain generalization that relies solely on positive data pairs, improving robustness without negative pair sampling.

Contribution

It proposes a novel regularization method using only positive pairs and a class-specific domain perturbation layer to enhance domain generalization.

Findings

Achieves comparable performance to state-of-the-art methods on DomainBed benchmark.

Effectively applies mixup augmentation with only positive pairs.

Improves domain generalization without negative pair sampling.

Abstract

In general, an experimental environment for deep learning assumes that the training and the test dataset are sampled from the same distribution. However, in real-world situations, a difference in the distribution between two datasets, domain shift, may occur, which becomes a major factor impeding the generalization performance of the model. The research field to solve this problem is called domain generalization, and it alleviates the domain shift problem by extracting domain-invariant features explicitly or implicitly. In recent studies, contrastive learning-based domain generalization approaches have been proposed and achieved high performance. These approaches require sampling of the negative data pair. However, the performance of contrastive learning fundamentally depends on quality and quantity of negative data pairs. To address this issue, we propose a new regularization method for domain generalization based on contrastive learning, self-supervised contrastive regularization (SelfReg). The proposed approach use only positive data pairs, thus it resolves various problems caused by negative pair sampling. Moreover, we propose a class-specific domain perturbation layer (CDPL), which makes it possible to effectively apply mixup augmentation even when only positive data pairs are used. The experimental results show that the techniques incorporated by SelfReg contributed to the performance in a compatible manner. In the recent benchmark, DomainBed, the proposed method shows comparable performance to the conventional state-of-the-art alternatives. Codes are available at https://github.com/dnap512/SelfReg.