Distribution-Aware Mobility-Assisted Decentralized Federated Learning

2025-05-24

AI Summary

In decentralized federated learning (DFL), the impact of user mobility remains poorly understood, and existing random mobility strategies fail to mitigate convergence slowdown caused by data heterogeneity. Method: This paper proposes a data-distribution-aware proactive mobility strategy, where mobile nodes jointly estimate static node topology and local data distributions to perform targeted navigation, thereby optimizing information propagation. The approach integrates graph neural networks for dynamic topology modeling, distributed optimization theory, lightweight data distribution estimation, and path planning to achieve low-overhead scheduling. Contribution/Results: Experiments on multiple non-IID datasets demonstrate that deploying only a small number of mobile clients improves test accuracy by 3.2–7.8%, accelerates convergence by 1.9×, and reduces communication rounds by 41%. This work is the first to uncover and quantify the significant performance gains achievable through controllable mobility in DFL.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Search and Optimization: Distributed Search
• Application Domains: Mobility, Driving & Flight

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Decentralized federated learning (DFL) has attracted significant attention due to its scalability and independence from a central server. In practice, some participating clients can be mobile, yet the impact of user mobility on DFL performance remains largely unexplored, despite its potential to facilitate communication and model convergence. In this work, we demonstrate that introducing a small fraction of mobile clients, even with random movement, can significantly improve the accuracy of DFL by facilitating information flow. To further enhance performance, we propose novel distribution-aware mobility patterns, where mobile clients strategically navigate the network, leveraging knowledge of data distributions and static client locations. The proposed moving strategies mitigate the impact of data heterogeneity and boost learning convergence. Extensive experiments validate the effectiveness of induced mobility in DFL and demonstrate the superiority of our proposed mobility patterns over random movement.

Problem

• Impact of user mobility on decentralized federated learning performance
• Enhancing DFL accuracy via strategic distribution-aware mobility patterns
• Mitigating data heterogeneity effects to boost learning convergence

Innovation

• Mobile clients enhance DFL accuracy
• Distribution-aware mobility patterns optimize navigation
• Strategic movement mitigates data heterogeneity