Summary
Synchronizing updates between devices and servers in edge federated learning is usually done on fixed schedules, which can slow down training especially when devices have different speeds and connection quality. The authors propose a method where each device decides on its own when and how to sync based on what it observes locally, like its learning progress and communication conditions. This locally adaptive approach helps improve the learning efficiency and accuracy in tasks like traffic prediction and image recognition while also considering privacy limits. The study shows that letting devices control synchronization is better than relying only on a central server schedule.
What this means in practice
- •For edge computing engineers: Implement adaptive client-side synchronization to improve energy use and latency in edge federated learning systems with varied device capabilities.
- •For mobile ai developers: Develop image classification apps that maintain accuracy while reducing communication costs by using adaptive synchronization during federated training.
Authors
Ebasa Temesgen, Ammar Ahmed, Nathan Lechamo, Ali Anwar, Maria Gini
Abstract
The scalability and performance of edge federated learning are fundamentally constrained by synchronization bottlenecks. Clients operate under heterogeneous compute, time-varying communication conditions, and, in some settings, privacy constraints, yet synchronization is often still handled with fixed schedules. However, synchronization decisions are often made outside the client, even though the client directly observes local progress, resource state, and communication conditions. We present a client-side reinforcement learning synchronization framework in which each client decides during training whether to upload, download, synchronize bidirectionally, or defer communication. The framework places a lightweight scheduler alongside the local model and formulates synchronization as a resource-aware control problem driven by local learning and system state. We study this framework on two representative edge workloads, spatiotemporal traffic forecasting and image classification. Across both settings, adaptive client-side synchronization improves the accuracy-cost trade-off, reducing system cost in traffic forecasting while maintaining competitive cost and improving final accuracy in image classification. We also consider privacy-constrained settings, where synchronization must account for not only latency and energy but also privacy expenditure. Overall, our results show that synchronization in edge federated learning is better treated as an adaptive client-side control problem than only as a server-side coordination mechanism.