Learning by Sorting: Self-supervised Learning with Group Ordering Constraints

TL;DR

This paper introduces GroCo, a self-supervised learning method that uses differentiable sorting constraints to improve representation learning by better ordering positive and negative pairs.

Contribution

The paper proposes a novel contrastive learning variation using differentiable sorting networks to enforce group ordering constraints, enhancing local neighborhood optimization.

Findings

Improved results over vanilla contrastive learning

Competitive performance in linear probing

Outperforms current methods in k-NN tasks

Abstract

Contrastive learning has become an important tool in learning representations from unlabeled data mainly relying on the idea of minimizing distance between positive data pairs, e.g., views from the same images, and maximizing distance between negative data pairs, e.g., views from different images. This paper proposes a new variation of the contrastive learning objective, Group Ordering Constraints (GroCo), that leverages the idea of sorting the distances of positive and negative pairs and computing the respective loss based on how many positive pairs have a larger distance than the negative pairs, and thus are not ordered correctly. To this end, the GroCo loss is based on differentiable sorting networks, which enable training with sorting supervision by matching a differentiable permutation matrix, which is produced by sorting a given set of scores, to a respective ground truth permutation matrix. Applying this idea to groupwise pre-ordered inputs of multiple positive and negative pairs allows introducing the GroCo loss with implicit emphasis on strong positives and negatives, leading to better optimization of the local neighborhood. We evaluate the proposed formulation on various self-supervised learning benchmarks and show that it not only leads to improved results compared to vanilla contrastive learning but also shows competitive performance to comparable methods in linear probing and outperforms current methods in k-NN performance.