ProxyFL: A Proxy-Guided Framework for Federated Semi-Supervised Learning

2026-02-24

AI Summary

This work addresses the challenges of both inter-client (external) and intra-client (internal) data heterogeneity in federated semi-supervised learning, particularly the mismatch between labeled and unlabeled data distributions and the bias introduced during global model aggregation. To tackle these issues, the authors propose ProxyFL, a novel framework that introduces a unified proxy mechanism leveraging learnable classifier weights as proxies for class distributions. The method mitigates external heterogeneity by optimizing a global proxy and combats internal heterogeneity through a positive-negative proxy pool that reuses discarded samples. ProxyFL further integrates pseudo-label correction and proxy-guided optimization strategies. Experimental results demonstrate that ProxyFL significantly improves model performance and convergence speed across multiple benchmarks, effectively overcoming the adverse effects of data heterogeneity.

Technology Category

• Machine Learning: Semi-Supervised Learning
• Search and Optimization: Learning to Search
• Multiagent Systems: Multiagent Learning

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• 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

Abstract

Federated Semi-Supervised Learning (FSSL) aims to collaboratively train a global model across clients by leveraging partially-annotated local data in a privacy-preserving manner. In FSSL, data heterogeneity is a challenging issue, which exists both across clients and within clients. External heterogeneity refers to the data distribution discrepancy across different clients, while internal heterogeneity represents the mismatch between labeled and unlabeled data within clients. Most FSSL methods typically design fixed or dynamic parameter aggregation strategies to collect client knowledge on the server (external) and / or filter out low-confidence unlabeled samples to reduce mistakes in local client (internal). But, the former is hard to precisely fit the ideal global distribution via direct weights, and the latter results in fewer data participation into FL training. To this end, we propose a proxy-guided framework called ProxyFL that focuses on simultaneously mitigating external and internal heterogeneity via a unified proxy. I.e., we consider the learnable weights of classifier as proxy to simulate the category distribution both locally and globally. For external, we explicitly optimize global proxy against outliers instead of direct weights; for internal, we re-include the discarded samples into training by a positive-negative proxy pool to mitigate the impact of potentially-incorrect pseudo-labels. Insight experiments&theoretical analysis show our significant performance and convergence in FSSL.

Problem

• Federated Semi-Supervised Learning
• Data Heterogeneity
• External Heterogeneity
• Internal Heterogeneity
• Pseudo-labels

Innovation

• Proxy-Guided Learning
• Federated Semi-Supervised Learning
• Data Heterogeneity
• Classifier Proxy
• Pseudo-Label Refinement