Oct 2026· IEEE Transactions on Parallel and Distributed Systems· Vol 37, pp. 2268-2282· 1 citation· 38 references
Computer Science
TL;DR
This work proposes FedASRA, an efficient semi-asynchronous federated learning framework featuring adaptive alternating selection (AAS) and real-time staleness adjustment (RSA), and provides convergence and system complexity analysis for FedASRA, establishing a staleness-agnostic convergence upper bound.
Abstract
Asynchronous federated learning is a distributed machine learning framework designed for parallel and distributed systems. However, system heterogeneity in communication and computation can cause stragglers to upload stale updates, which may degrade the performance of the global model. Existing studies primarily focus on systems with mild to moderate heterogeneity, leaving the trade-off between communication efficiency and model performance insufficiently addressed in extreme scenarios. To tackle this issue, we propose FedASRA, an efficient semi-asynchronous federated learning framework featuring adaptive alternating selection (AAS) and real-time staleness adjustment (RSA). Specifically, AAS leverages a multi-attribute utility function to periodically schedule the participation of stragglers and fast clients in an alternating manner, thereby narrowing the system heterogeneity gap among clients in each communication round. When staleness is unavoidable, RSA enables stragglers to synchronously receive the latest global model during training and compute updated gradients. These updated gradients are then locally aggregated with the soon-to-be stale gradients to improve the freshness of submitted updates. We provide convergence and system complexity analysis for FedASRA, establishing a staleness-agnostic convergence upper bound. Extensive experiments are conducted in simulated environments and a prototype system built with 21 real-world devices. Results demonstrate that FedASRA achieves a favorable balance between system efficiency and performance. Compared with state-of-the-art methods, FedASRA reduces wall-clock time by 18.51% and communication overhead by 11.66% on average.
Federated learning enables multiple clients to collaboratively train a shared machine-learning model without directly poexchanging their raw data. In distributed implementations, Ring-All-Reduce is commonly used to aggregate model updates efficiently, but its standard aggregation operations can be vulnerable when one o...
D. S. K. Sathe Rohit Adinath, B. H. J. N. M. Kolhe· International Journal of Adv...· 0 citations
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. Howeve...
Ebasa Temesgen, Ammar Ahmed, N. Lechamo et al.· 0 citations
In this paper, we introduce FLEX, a flexible federated learning (FL) scheme that combines peer-to-peer communication with asynchronous client updates, enabling clients to operate independently. We further propose a novel aggregation algorithm inspired by gradient-like updates, governed by two key parameters: (i) a simi...
C. Sad, D. Masouros, George Retsinas et al.· IEEE Transactions on Paralle...· 0 citations