DimCL: Dimensional Contrastive Learning For Improving Self-Supervised Learning

TL;DR

Dimensional Contrastive Learning (DimCL) introduces a novel contrastive approach along feature dimensions, enhancing feature diversity and serving as a regularizer to improve self-supervised learning performance across multiple datasets.

Contribution

The paper proposes DimCL, a new contrastive learning strategy along feature dimensions, which effectively improves SSL frameworks and explains its success through hardness-aware properties.

Findings

DimCL improves SSL performance on various datasets.

DimCL enhances feature diversity and acts as a regularizer.

Extensive experiments confirm the effectiveness of DimCL.

Abstract

Self-supervised learning (SSL) has gained remarkable success, for which contrastive learning (CL) plays a key role. However, the recent development of new non-CL frameworks has achieved comparable or better performance with high improvement potential, prompting researchers to enhance these frameworks further. Assimilating CL into non-CL frameworks has been thought to be beneficial, but empirical evidence indicates no visible improvements. In view of that, this paper proposes a strategy of performing CL along the dimensional direction instead of along the batch direction as done in conventional contrastive learning, named Dimensional Contrastive Learning (DimCL). DimCL aims to enhance the feature diversity, and it can serve as a regularizer to prior SSL frameworks. DimCL has been found to be effective, and the hardness-aware property is identified as a critical reason for its success. Extensive experimental results reveal that assimilating DimCL into SSL frameworks leads to performance improvement by a non-trivial margin on various datasets and backbone architectures.