Bayesian Self-Supervised Contrastive Learning

TL;DR

This paper introduces the Bayesian Contrastive Loss (BCL) for self-supervised contrastive learning, which corrects sampling bias using importance weights and a Bayesian framework to improve negative sample selection.

Contribution

It proposes a novel BCL loss that models the sampling distribution parametrically to effectively handle false negatives and hard negatives in self-supervised contrastive learning.

Findings

BCL outperforms existing contrastive loss methods in experiments.

The Bayesian framework effectively reduces false negatives.

Parametric sampling improves hard negative mining.

Abstract

Recent years have witnessed many successful applications of contrastive learning in diverse domains, yet its self-supervised version still remains many exciting challenges. As the negative samples are drawn from unlabeled datasets, a randomly selected sample may be actually a false negative to an anchor, leading to incorrect encoder training. This paper proposes a new self-supervised contrastive loss called the BCL loss that still uses random samples from the unlabeled data while correcting the resulting bias with importance weights. The key idea is to design the desired sampling distribution for sampling hard true negative samples under the Bayesian framework. The prominent advantage lies in that the desired sampling distribution is a parametric structure, with a location parameter for debiasing false negative and concentration parameter for mining hard negative, respectively. Experiments validate the effectiveness and superiority of the BCL loss.