Parametric Contrastive Learning

TL;DR

Parametric Contrastive Learning (PaCo) introduces class-wise learnable centers to improve long-tailed recognition, outperforming existing methods by adaptively balancing class representations and enhancing hard example learning.

Contribution

The paper proposes a novel parametric contrastive loss with class-wise centers, addressing bias in supervised contrastive learning for imbalanced data.

Findings

Achieves state-of-the-art results on long-tailed datasets.

Surpasses supervised contrastive learning on full ImageNet.

Effectively balances classes and improves hard example learning.

Abstract

In this paper, we propose Parametric Contrastive Learning (PaCo) to tackle long-tailed recognition. Based on theoretical analysis, we observe supervised contrastive loss tends to bias on high-frequency classes and thus increases the difficulty of imbalanced learning. We introduce a set of parametric class-wise learnable centers to rebalance from an optimization perspective. Further, we analyze our PaCo loss under a balanced setting. Our analysis demonstrates that PaCo can adaptively enhance the intensity of pushing samples of the same class close as more samples are pulled together with their corresponding centers and benefit hard example learning. Experiments on long-tailed CIFAR, ImageNet, Places, and iNaturalist 2018 manifest the new state-of-the-art for long-tailed recognition. On full ImageNet, models trained with PaCo loss surpass supervised contrastive learning across various ResNet backbones, e.g., our ResNet-200 achieves 81.8% top-1 accuracy. Our code is available at https://github.com/dvlab-research/Parametric-Contrastive-Learning.