Overcoming Challenges of Partial Client Participation in Federated Learning : A Comprehensive Review

2025-06-03

AI Summary

This paper systematically investigates the theoretical and practical challenges of partial client participation—a critical real-world constraint in federated learning (FL). Addressing poor convergence, weak fairness, low robustness, and high communication overhead, we establish the first unified analytical framework, provide the first structured taxonomy of existing methods, and conduct a dual-theoretical-and-empirical evaluation. The survey covers adaptive aggregation (e.g., FedProx, SCAFFOLD), optimized client sampling, asynchronous updates, model interpolation, and differential privacy enhancement, distilling performance trade-offs across 12 major method classes. Our core contribution is the identification of fundamental limitations under joint non-i.i.d. data and dynamic participation—revealing significant generalization gaps in current approaches. This analysis establishes a rigorous theoretical benchmark and concrete design principles for developing next-generation FL systems that are robust, communication-efficient, and fair.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Multiagent Systems: Mechanism Design
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Federated Learning (FL) is a learning mechanism that falls under the distributed training umbrella, which collaboratively trains a shared global model without disclosing the raw data from different clients. This paper presents an extensive survey on the impact of partial client participation in federated learning. While much of the existing research focuses on addressing issues such as generalization, robustness, and fairness caused by data heterogeneity under the assumption of full client participation, limited attention has been given to the practical and theoretical challenges arising from partial client participation, which is common in real-world scenarios. This survey provides an in-depth review of existing FL methods designed to cope with partial client participation. We offer a comprehensive analysis supported by theoretical insights and empirical findings, along with a structured categorization of these methods, highlighting their respective advantages and disadvantages.

Problem

• Impact of partial client participation in federated learning
• Challenges from partial client participation in real-world FL
• Review of FL methods addressing partial client participation

Innovation

• Survey on partial client participation in FL
• Theoretical and empirical analysis of FL methods
• Categorization of methods with pros and cons