Contrastive Learning with Boosted Memorization

TL;DR

This paper introduces Boosted Contrastive Learning (BCL), a novel method leveraging neural network memorization to improve long-tailed data representation in self-supervised learning without labels.

Contribution

It proposes a data-centric approach that uses memorization effects to enhance contrastive learning for long-tailed distributions, outperforming existing methods.

Findings

BCL outperforms several state-of-the-art methods on benchmark datasets.

The memorization-based approach effectively discovers tail samples.

BCL demonstrates robustness across different long-tailed scenarios.

Abstract

Self-supervised learning has achieved a great success in the representation learning of visual and textual data. However, the current methods are mainly validated on the well-curated datasets, which do not exhibit the real-world long-tailed distribution. Recent attempts to consider self-supervised long-tailed learning are made by rebalancing in the loss perspective or the model perspective, resembling the paradigms in the supervised long-tailed learning. Nevertheless, without the aid of labels, these explorations have not shown the expected significant promise due to the limitation in tail sample discovery or the heuristic structure design. Different from previous works, we explore this direction from an alternative perspective, i.e., the data perspective, and propose a novel Boosted Contrastive Learning (BCL) method. Specifically, BCL leverages the memorization effect of deep neural networks to automatically drive the information discrepancy of the sample views in contrastive learning, which is more efficient to enhance the long-tailed learning in the label-unaware context. Extensive experiments on a range of benchmark datasets demonstrate the effectiveness of BCL over several state-of-the-art methods. Our code is available at https://github.com/MediaBrain-SJTU/BCL.