FedVG: Gradient-Guided Aggregation for Enhanced Federated Learning

TL;DR

FedVG introduces a gradient-guided aggregation method for federated learning that uses a global validation set to improve model generalization across heterogeneous clients.

Contribution

The paper proposes FedVG, a novel gradient-based aggregation framework that assesses client models using validation gradients, enhancing federated learning performance in diverse data settings.

Findings

FedVG consistently improves model accuracy in heterogeneous federated learning scenarios.

It can be integrated with existing FL algorithms to boost their performance.

Experiments on natural and medical images validate its effectiveness.

Abstract

Federated Learning (FL) enables collaborative model training across multiple clients without sharing their private data. However, data heterogeneity across clients leads to client drift, which degrades the overall generalization performance of the model. This effect is further compounded by overemphasis on poorly performing clients. To address this problem, we propose FedVG, a novel gradient-based federated aggregation framework that leverages a global validation set to guide the optimization process. Such a global validation set can be established using readily available public datasets, ensuring accessibility and consistency across clients without compromising privacy. In contrast to conventional approaches that prioritize client dataset volume, FedVG assesses the generalization ability of client models by measuring the magnitude of validation gradients across layers. Specifically, we compute layerwise gradient norms to derive a client-specific score that reflects how much each client needs to adjust for improved generalization on the global validation set, thereby enabling more informed and adaptive federated aggregation. Extensive experiments on both natural and medical image benchmarking datasets, across diverse model architectures, demonstrate that FedVG consistently improves performance, particularly in highly heterogeneous settings. Moreover, FedVG is modular and can be seamlessly integrated with various state-of-the-art FL algorithms, often further improving their results. Our code is available at https://github.com/alinadevkota/FedVG.