Mobility-Aware Asynchronous Federated Learning with Dynamic Sparsification

2025-06-08

AI Summary

In asynchronous federated learning (AFL) on mobile devices, device mobility induces intermittent connectivity, gradient sparsification, and model staleness—three interdependent challenges that jointly impede convergence. This paper presents the first theoretical convergence framework jointly modeling these three factors. We propose Mobility-Aware Dynamic Sparsification (MADS), a novel sparsification strategy that adaptively adjusts the sparsity level based on device contact duration and model staleness, and derive its closed-form optimal solution. Evaluated on CIFAR-10 image classification, MADS improves test accuracy by 8.76%; on Argoverse trajectory prediction, it reduces mean displacement error by 9.46%. These results demonstrate substantial improvements in both convergence behavior and practical effectiveness of AFL under realistic mobile network conditions.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Application Domains: Mobility, Driving & Flight
• Multiagent Systems: Adversarial Agents

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Asynchronous Federated Learning (AFL) enables distributed model training across multiple mobile devices, allowing each device to independently update its local model without waiting for others. However, device mobility introduces intermittent connectivity, which necessitates gradient sparsification and leads to model staleness, jointly affecting AFL convergence. This paper develops a theoretical model to characterize the interplay among sparsification, model staleness and mobility-induced contact patterns, and their joint impact on AFL convergence. Based on the analysis, we propose a mobility-aware dynamic sparsification (MADS) algorithm that optimizes the sparsification degree based on contact time and model staleness. Closed-form solutions are derived, showing that under low-speed conditions, MADS increases the sparsification degree to enhance convergence, while under high-speed conditions, it reduces the sparsification degree to guarantee reliable uploads within limited contact time. Experimental results validate the theoretical findings. Compared with the state-of-the-art benchmarks, the MADS algorithm increases the image classification accuracy on the CIFAR-10 dataset by 8.76% and reduces the average displacement error in the Argoverse trajectory prediction dataset by 9.46%.

Problem

• Analyzes impact of mobility on AFL convergence with sparsification
• Proposes MADS algorithm to optimize sparsification for mobility conditions
• Improves accuracy in image classification and trajectory prediction tasks

Innovation

• Dynamic sparsification optimizes AFL convergence
• Mobility-aware algorithm adjusts sparsification degree
• Closed-form solutions balance speed and reliability