Hyperspherical Consistency Regularization

TL;DR

This paper introduces Hyperspherical Consistency Regularization (HCR), a novel regularization technique for classifiers in semi-supervised learning that reduces bias and improves data efficiency by enforcing feature similarity on a hypersphere.

Contribution

HCR is a simple, plug-and-play regularization method that projects logits and features onto a hypersphere to promote consistent data structures, addressing classifier bias in semi-supervised learning.

Findings

HCR improves performance in semi-supervised and weakly-supervised tasks.

HCR effectively reduces classifier bias and enhances data efficiency.

Extensive experiments validate the superiority of HCR over existing methods.

Abstract

Recent advances in contrastive learning have enlightened diverse applications across various semi-supervised fields. Jointly training supervised learning and unsupervised learning with a shared feature encoder becomes a common scheme. Though it benefits from taking advantage of both feature-dependent information from self-supervised learning and label-dependent information from supervised learning, this scheme remains suffering from bias of the classifier. In this work, we systematically explore the relationship between self-supervised learning and supervised learning, and study how self-supervised learning helps robust data-efficient deep learning. We propose hyperspherical consistency regularization (HCR), a simple yet effective plug-and-play method, to regularize the classifier using feature-dependent information and thus avoid bias from labels. Specifically, HCR first projects logits from the classifier and feature projections from the projection head on the respective hypersphere, then it enforces data points on hyperspheres to have similar structures by minimizing binary cross entropy of pairwise distances' similarity metrics. Extensive experiments on semi-supervised and weakly-supervised learning demonstrate the effectiveness of our method, by showing superior performance with HCR.