Contrastive Learning with Consistent Representations

TL;DR

This paper introduces CoCor, a contrastive learning framework that uses a novel DA consistency metric to improve representation quality by systematically managing the effects of data augmentation.

Contribution

The paper proposes a new consistency metric and a method to learn optimal representation mappings conditioned on data augmentation, enhancing contrastive learning robustness.

Findings

Improved generalization of learned representations.

Enhanced transferability across tasks.

Systematic management of data augmentation effects.

Abstract

Contrastive learning demonstrates great promise for representation learning. Data augmentations play a critical role in contrastive learning by providing informative views of the data without necessitating explicit labels. Nonetheless, the efficacy of current methodologies heavily hinges on the quality of employed data augmentation (DA) functions, often chosen manually from a limited set of options. While exploiting diverse data augmentations is appealing, the complexities inherent in both DAs and representation learning can lead to performance deterioration. Addressing this challenge and facilitating the systematic incorporation of diverse data augmentations, this paper proposes Contrastive Learning with Consistent Representations CoCor. At the heart of CoCor is a novel consistency metric termed DA consistency. This metric governs the mapping of augmented input data to the representation space, ensuring that these instances are positioned optimally in a manner consistent with the applied intensity of the DA. Moreover, we propose to learn the optimal mapping locations as a function of DA, all while preserving a desired monotonic property relative to DA intensity. Experimental results demonstrate that CoCor notably enhances the generalizability and transferability of learned representations in comparison to baseline methods.