Convergence of Agnostic Federated Averaging

2025-07-14

AI Summary

This paper addresses the convergence of Agnostic FedAvg under realistic client participation patterns in federated learning—specifically, random, non-uniform, and distribution-agnostic client availability. Existing analyses typically assume either full-client participation or known uniform availability distributions, which severely diverge from practical deployment scenarios. Method: We propose the first rigorous convergence analysis of Agnostic FedAvg for non-uniform stochastic participation without prior knowledge of client sampling probabilities, modeling the dynamic aggregation process under convex (including nonsmooth) loss functions. Results: We establish a standard $O(1/sqrt{T})$ convergence rate and theoretically prove that our bound is tighter than those of weighted aggregation methods relying on known participation weights. Extensive experiments confirm consistent superiority over conventional variants across diverse heterogeneous participation patterns. Our core contribution lies in eliminating reliance on prior knowledge of participation distributions, thereby providing the first unbiased and robust convergence guarantee for real-world federated learning deployments.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Reasoning under Uncertainty: Stochastic Optimization
• Search and Optimization: Non-convex Optimization

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
• Security and Privacy: Large-scale security measurements

Abstract

Federated learning (FL) enables decentralized model training without centralizing raw data. However, practical FL deployments often face a key realistic challenge: Clients participate intermittently in server aggregation and with unknown, possibly biased participation probabilities. Most existing convergence results either assume full-device participation, or rely on knowledge of (in fact uniform) client availability distributions -- assumptions that rarely hold in practice. In this work, we characterize the optimization problem that consistently adheres to the stochastic dynamics of the well-known emph{agnostic Federated Averaging (FedAvg)} algorithm under random (and variably-sized) client availability, and rigorously establish its convergence for convex, possibly nonsmooth losses, achieving a standard rate of order $mathcal{O}(1/sqrt{T})$, where $T$ denotes the aggregation horizon. Our analysis provides the first convergence guarantees for agnostic FedAvg under general, non-uniform, stochastic client participation, without knowledge of the participation distribution. We also empirically demonstrate that agnostic FedAvg in fact outperforms common (and suboptimal) weighted aggregation FedAvg variants, even with server-side knowledge of participation weights.

Problem

• Analyzes convergence of FedAvg under random client participation
• Addresses unknown biased client availability in federated learning
• Proves convergence without knowing participation distribution

Innovation

• Agnostic Federated Averaging under random client availability
• Convergence guarantees without participation distribution knowledge
• Outperforms weighted aggregation FedAvg variants empirically