Fed-REACT: Federated Representation Learning for Heterogeneous and Evolving Data

2025-09-08

AI Summary

To address performance degradation in federated learning caused by heterogeneous and dynamically evolving client data distributions, this paper proposes a two-stage collaborative framework: clients locally learn robust feature representations, while the server performs evolutionary clustering in the representation space and organizes cluster-wise collaborative training accordingly. This work is the first to deeply integrate federated representation learning with dynamic evolutionary clustering, enabling on-demand clustering and personalized modeling. Theoretical analysis guarantees both clustering stability and algorithmic convergence. Extensive experiments on multiple real-world datasets demonstrate that the proposed method significantly outperforms state-of-the-art federated learning algorithms, achieving consistent improvements in classification accuracy, model robustness, and adaptability to concept drift.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Representation Learning for Vision
• Search and Optimization: Evolutionary Computation

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing

Abstract

Motivated by the high resource costs and privacy concerns associated with centralized machine learning, federated learning (FL) has emerged as an efficient alternative that enables clients to collaboratively train a global model while keeping their data local. However, in real-world deployments, client data distributions often evolve over time and differ significantly across clients, introducing heterogeneity that degrades the performance of standard FL algorithms. In this work, we introduce Fed-REACT, a federated learning framework designed for heterogeneous and evolving client data. Fed-REACT combines representation learning with evolutionary clustering in a two-stage process: (1) in the first stage, each client learns a local model to extracts feature representations from its data; (2) in the second stage, the server dynamically groups clients into clusters based on these representations and coordinates cluster-wise training of task-specific models for downstream objectives such as classification or regression. We provide a theoretical analysis of the representation learning stage, and empirically demonstrate that Fed-REACT achieves superior accuracy and robustness on real-world datasets.

Problem

• Addresses federated learning challenges with heterogeneous evolving data
• Combats performance degradation from non-IID client data distributions
• Enhances model accuracy and robustness in dynamic environments

Innovation

• Federated learning with representation learning
• Dynamic client clustering using evolutionary methods
• Two-stage process for heterogeneous data