Rethinking the Augmentation Module in Contrastive Learning: Learning Hierarchical Augmentation Invariance with Expanded Views

TL;DR

This paper introduces a novel contrastive learning framework that adaptively learns hierarchical augmentation invariances and expands contrast content with augmentation embeddings, improving representation quality across multiple tasks.

Contribution

It proposes a method to learn different augmentation invariances at various model depths and to expand contrast content using augmentation embeddings, addressing limitations of fixed augmentation strategies.

Findings

Improved representations for classification, detection, and segmentation tasks.

Enhanced flexibility in choosing augmentation types for different downstream tasks.

Better downstream task performance compared to baseline contrastive methods.

Abstract

A data augmentation module is utilized in contrastive learning to transform the given data example into two views, which is considered essential and irreplaceable. However, the predetermined composition of multiple data augmentations brings two drawbacks. First, the artificial choice of augmentation types brings specific representational invariances to the model, which have different degrees of positive and negative effects on different downstream tasks. Treating each type of augmentation equally during training makes the model learn non-optimal representations for various downstream tasks and limits the flexibility to choose augmentation types beforehand. Second, the strong data augmentations used in classic contrastive learning methods may bring too much invariance in some cases, and fine-grained information that is essential to some downstream tasks may be lost. This paper proposes a general method to alleviate these two problems by considering where and what to contrast in a general contrastive learning framework. We first propose to learn different augmentation invariances at different depths of the model according to the importance of each data augmentation instead of learning representational invariances evenly in the backbone. We then propose to expand the contrast content with augmentation embeddings to reduce the misleading effects of strong data augmentations. Experiments based on several baseline methods demonstrate that we learn better representations for various benchmarks on classification, detection, and segmentation downstream tasks.