EPFL-REMNet: Efficient Personalized Federated Digital Twin Towards 6G Heterogeneous Radio Environment

2025-11-07

AI Summary

To address the degradation in federated learning (FL) accuracy and high uplink communication overhead caused by non-independent and identically distributed (Non-IID) data in 6G heterogeneous wireless environments, this paper proposes a lightweight personalized FL framework tailored for high-fidelity radio-frequency digital twin map construction. The core innovation is a “shared backbone + lightweight personalized head” architecture: only compressed global backbone parameters are uploaded to the server, while each client retains a compact, locally adapted head for personalization; this is further enhanced by geography-aware data partitioning and model compression. Evaluations across 90 clients under three distinct Non-IID settings demonstrate that the proposed method significantly improves digital twin fidelity (average gain of 12.7%), reduces uplink communication overhead by 63.5%, and substantially enhances local map accuracy for tail-end clients (RMSE reduction of 28.4%).

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Search and Optimization: Learning to Search
• Natural Language Processing: Language Grounding & Multi-modal NLP

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Radio Environment Map (REM) is transitioning from 5G homogeneous environments to B5G/6G heterogeneous landscapes. However, standard Federated Learning (FL), a natural fit for this distributed task, struggles with performance degradation in accuracy and communication efficiency under the non-independent and identically distributed (Non-IID) data conditions inherent to these new environments. This paper proposes EPFL-REMNet, an efficient personalized federated framework for constructing a high-fidelity digital twin of the 6G heterogeneous radio environment. The proposed EPFL-REMNet employs a"shared backbone + lightweight personalized head"model, where only the compressed shared backbone is transmitted between the server and clients, while each client's personalized head is maintained locally. We tested EPFL-REMNet by constructing three distinct Non-IID scenarios (light, medium, and heavy) based on radio environment complexity, with data geographically partitioned across 90 clients. Experimental results demonstrate that EPFL-REMNet simultaneously achieves higher digital twin fidelity (accuracy) and lower uplink overhead across all Non-IID settings compared to standard FedAvg and recent state-of-the-art methods. Particularly, it significantly reduces performance disparities across datasets and improves local map accuracy for long-tail clients, enhancing the overall integrity of digital twin.

Problem

• Addresses performance degradation in federated learning under Non-IID data conditions
• Improves digital twin fidelity and communication efficiency for 6G radio environments
• Reduces performance disparities across datasets and enhances local map accuracy

Innovation

• Shared backbone with lightweight personalized head model
• Compressed shared backbone transmission between server clients
• Local maintenance of personalized heads for each client