FedeCouple: Fine-Grained Balancing of Global-Generalization and Local-Adaptability in Federated Learning

TL;DR

FedeCouple introduces a novel federated learning approach that finely balances global model generalization with local client adaptability, improving personalized model performance while preserving privacy and reducing communication costs.

Contribution

It proposes a new method that jointly learns global and local features with dynamic knowledge distillation and anchor-based feature refinement, ensuring convergence and superior performance.

Findings

Outperforms nine baseline methods in effectiveness, stability, scalability, and security.

Achieves a 4.3% improvement over the best baseline in effectiveness.

Proven convergence for nonconvex objectives with theoretical analysis.

Abstract

In privacy-preserving mobile network transmission scenarios with heterogeneous client data, personalized federated learning methods that decouple feature extractors and classifiers have demonstrated notable advantages in enhancing learning capability. However, many existing approaches primarily focus on feature space consistency and classification personalization during local training, often neglecting the local adaptability of the extractor and the global generalization of the classifier. This oversight results in insufficient coordination and weak coupling between the components, ultimately degrading the overall model performance. To address this challenge, we propose FedeCouple, a federated learning method that balances global generalization and local adaptability at a fine-grained level. Our approach jointly learns global and local feature representations while employing dynamic knowledge distillation to enhance the generalization of personalized classifiers. We further introduce anchors to refine the feature space; their strict locality and non-transmission inherently preserve privacy and reduce communication overhead. Furthermore, we provide a theoretical analysis proving that FedeCouple converges for nonconvex objectives, with iterates approaching a stationary point as the number of communication rounds increases. Extensive experiments conducted on five image-classification datasets demonstrate that FedeCouple consistently outperforms nine baseline methods in effectiveness, stability, scalability, and security. Notably, in experiments evaluating effectiveness, FedeCouple surpasses the best baseline by a significant margin of 4.3%.