Contrastive Learning with Adversarial Examples

TL;DR

This paper introduces CLAE, an adversarial training method for contrastive learning that generates challenging negative pairs, leading to improved visual representation learning across multiple datasets.

Contribution

It proposes a novel adversarial example generation technique and an adversarial training algorithm for contrastive learning, enhancing the difficulty of negative pairs and improving SSL performance.

Findings

CLAE outperforms standard contrastive learning baselines.

Adversarial examples improve the difficulty of negative pairs.

Enhanced SSL representations on multiple datasets.

Abstract

Contrastive learning (CL) is a popular technique for self-supervised learning (SSL) of visual representations. It uses pairs of augmentations of unlabeled training examples to define a classification task for pretext learning of a deep embedding. Despite extensive works in augmentation procedures, prior works do not address the selection of challenging negative pairs, as images within a sampled batch are treated independently. This paper addresses the problem, by introducing a new family of adversarial examples for constrastive learning and using these examples to define a new adversarial training algorithm for SSL, denoted as CLAE. When compared to standard CL, the use of adversarial examples creates more challenging positive pairs and adversarial training produces harder negative pairs by accounting for all images in a batch during the optimization. CLAE is compatible with many CL methods in the literature. Experiments show that it improves the performance of several existing CL baselines on multiple datasets.