AFBS:Buffer Gradient Selection in Semi-asynchronous Federated Learning

TL;DR

AFBS introduces a novel gradient selection method in semi-asynchronous federated learning, improving accuracy and efficiency by discarding low-value gradients while maintaining privacy, especially in heterogeneous environments.

Contribution

This paper presents AFBS, the first algorithm to perform gradient selection within buffers in federated learning, enhancing performance and privacy protection.

Findings

AFBS outperforms state-of-the-art methods in heterogeneous settings.

On CIFAR-100, AFBS improves accuracy by up to 4.8%.

It reduces time to target accuracy by 75%.

Abstract

Asynchronous federated learning (AFL) accelerates training by eliminating the need to wait for stragglers, but its asynchronous nature introduces gradient staleness, where outdated gradients degrade performance. Existing solutions address this issue with gradient buffers, forming a semi-asynchronous framework. However, this approach struggles when buffers accumulate numerous stale gradients, as blindly aggregating all gradients can harm training. To address this, we propose AFBS (Asynchronous FL Buffer Selection), the first algorithm to perform gradient selection within buffers while ensuring privacy protection. Specifically, the client sends the random projection encrypted label distribution matrix before training, and the server performs client clustering based on it. During training, server scores and selects gradients within each cluster based on their informational value, discarding low-value gradients to enhance semi-asynchronous federated learning. Extensive experiments in highly heterogeneous system and data environments demonstrate AFBS's superior performance compared to state-of-the-art methods. Notably, on the most challenging task, CIFAR-100, AFBS improves accuracy by up to 4.8% over the previous best algorithm and reduces the time to reach target accuracy by 75%.