Class-Aware Contrastive Semi-Supervised Learning

TL;DR

This paper introduces Class-aware Contrastive Semi-Supervised Learning (CCSSL), a method that improves pseudo-label quality and robustness in semi-supervised learning by handling in-distribution and out-of-distribution data separately with class-wise clustering and contrastive learning.

Contribution

The paper proposes a novel CCSSL method that enhances semi-supervised learning by combining class-aware clustering and contrastive techniques to better handle noisy and out-of-distribution data.

Findings

Significant performance gains on CIFAR100 and STL10 datasets.

Improves FixMatch accuracy by 9.80% on Semi-iNat 2021.

Enhances robustness and reduces confirmation bias in SSL.

Abstract

Pseudo-label-based semi-supervised learning (SSL) has achieved great success on raw data utilization. However, its training procedure suffers from confirmation bias due to the noise contained in self-generated artificial labels. Moreover, the model's judgment becomes noisier in real-world applications with extensive out-of-distribution data. To address this issue, we propose a general method named Class-aware Contrastive Semi-Supervised Learning (CCSSL), which is a drop-in helper to improve the pseudo-label quality and enhance the model's robustness in the real-world setting. Rather than treating real-world data as a union set, our method separately handles reliable in-distribution data with class-wise clustering for blending into downstream tasks and noisy out-of-distribution data with image-wise contrastive for better generalization. Furthermore, by applying target re-weighting, we successfully emphasize clean label learning and simultaneously reduce noisy label learning. Despite its simplicity, our proposed CCSSL has significant performance improvements over the state-of-the-art SSL methods on the standard datasets CIFAR100 and STL10. On the real-world dataset Semi-iNat 2021, we improve FixMatch by 9.80% and CoMatch by 3.18%. Code is available https://github.com/TencentYoutuResearch/Classification-SemiCLS.