Synthetic Hard Negative Samples for Contrastive Learning

TL;DR

This paper introduces SSCL, a novel method for generating and sampling synthetic hard negative samples at the feature level to enhance contrastive learning, leading to improved visual representation performance.

Contribution

The paper proposes a new feature-level sampling technique for synthetic hard negatives, improving contrastive learning effectiveness and addressing false negatives.

Findings

Enhanced classification accuracy on multiple image datasets

Effective integration with existing contrastive learning methods

Better utilization of hard negative samples

Abstract

Contrastive learning has emerged as an essential approach for self-supervised learning in visual representation learning. The central objective of contrastive learning is to maximize the similarities between two augmented versions of an image (positive pairs), while minimizing the similarities between different images (negative pairs). Recent studies have demonstrated that harder negative samples, i.e., those that are more difficult to differentiate from the anchor sample, perform a more crucial function in contrastive learning. This paper proposes a novel feature-level method, namely sampling synthetic hard negative samples for contrastive learning (SSCL), to exploit harder negative samples more effectively. Specifically, 1) we generate more and harder negative samples by mixing negative samples, and then sample them by controlling the contrast of anchor sample with the other negative samples; 2) considering the possibility of false negative samples, we further debias the negative samples. Our proposed method improves the classification performance on different image datasets and can be readily integrated into existing methods.