Leveraging Label Proportion Prior for Class-Imbalanced Semi-Supervised Learning

2026-03-03

AI Summary

This work addresses the challenge of class imbalance in semi-supervised learning, where pseudo-labeling often exacerbates bias toward majority classes and degrades performance on minority classes. To mitigate this issue, the authors propose the first integration of Proportion Loss—a technique originally developed in Learning from Label Proportions (LLP)—into semi-supervised frameworks, regularizing model predictions to align with the global class distribution. They further introduce a stochastic variant of this loss to accommodate the inherent instability of mini-batch training. Built upon FixMatch and ReMixMatch, the proposed approach consistently achieves state-of-the-art or competitive performance across varying degrees of imbalance and label proportions on long-tailed CIFAR-10, with particularly pronounced gains under extremely limited labeling conditions.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Search and Optimization: Learning to Search
• Computer Vision: Learning & Optimization for CV

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Semi-supervised learning (SSL) often suffers under class imbalance, where pseudo-labeling amplifies majority bias and suppresses minority performance. We address this issue with a lightweight framework that, to our knowledge, is the first to introduce Proportion Loss from learning from label proportions (LLP) into SSL as a regularization term. Proportion Loss aligns model predictions with the global class distribution, mitigating bias across both majority and minority classes. To further stabilize training, we formulate a stochastic variant that accounts for fluctuations in mini-batch composition. Experiments on the Long-tailed CIFAR-10 benchmark show that integrating Proportion Loss into FixMatch and ReMixMatch consistently improves performance over the baselines across imbalance severities and label ratios, and achieves competitive or superior results compared to existing CISSL methods, particularly under scarce-label conditions.

Problem

• class imbalance
• semi-supervised learning
• pseudo-labeling
• label proportion
• minority class

Innovation

• Proportion Loss
• Class-Imbalanced Semi-Supervised Learning
• Learning from Label Proportions
• FixMatch
• Long-tailed Recognition