Balancing Client Participation in Federated Learning Using AoI

2025-05-08

AI Summary

In federated learning, uneven client participation leads to unstable convergence and imbalanced communication loads. To address this, we propose the first Age-of-Information (AoI)-driven decentralized client selection mechanism. Our method introduces a distributed Markov scheduling policy that probabilistically selects clients based on their AoI—i.e., the elapsed time since their last participation—thereby achieving cross-round load balancing without centralized coordination. This ensures both participation fairness and convergence stability. We provide a rigorous theoretical convergence analysis under general non-convex objectives. Extensive experiments demonstrate that our approach significantly outperforms FedAvg under both IID and non-IID data settings, accelerating convergence by 7.5%–20%, while simultaneously improving model stability and system scalability.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Multiagent Systems: Mechanism Design
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Decentralized Web and Fediverse systems
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments

Abstract

Federated Learning (FL) offers a decentralized framework that preserves data privacy while enabling collaborative model training across distributed clients. However, FL faces significant challenges due to limited communication resources, statistical heterogeneity, and the need for balanced client participation. This paper proposes an Age of Information (AoI)-based client selection policy that addresses these challenges by minimizing load imbalance through controlled selection intervals. Our method employs a decentralized Markov scheduling policy, allowing clients to independently manage participation based on age-dependent selection probabilities, which balances client updates across training rounds with minimal central oversight. We provide a convergence proof for our method, demonstrating that it ensures stable and efficient model convergence. Specifically, we derive optimal parameters for the Markov selection model to achieve balanced and consistent client participation, highlighting the benefits of AoI in enhancing convergence stability. Through extensive simulations, we demonstrate that our AoI-based method, particularly the optimal Markov variant, improves convergence over the FedAvg selection approach across both IID and non-IID data settings by $7.5%$ and up to $20%$. Our findings underscore the effectiveness of AoI-based scheduling for scalable, fair, and efficient FL systems across diverse learning environments.

Problem

• Balances client participation in Federated Learning using AoI
• Addresses communication limits and data heterogeneity in FL
• Improves convergence stability with decentralized Markov scheduling

Innovation

• AoI-based client selection policy
• Decentralized Markov scheduling policy
• Optimal parameters for balanced participation