Efficient Federated Learning with Encrypted Data Sharing for Data-Heterogeneous Edge Devices

2025-06-25

AI Summary

To address high training latency, slow convergence, and degraded model performance in edge federated learning—caused by dynamic network topologies, physical distances among devices, and severe data heterogeneity—this paper proposes FedEDS, a novel framework. FedEDS introduces a lightweight, client-side data encryptor dynamically generated from stochastic layers of local models, enabling secure, low-overhead cross-device encrypted data sharing. This mechanism mitigates data heterogeneity while enhancing local training efficacy. Integrating edge-distributed architecture, federated collaborative optimization, and encrypted data augmentation, FedEDS improves model consistency without increasing communication overhead. Extensive experiments on multiple heterogeneous edge datasets demonstrate that FedEDS accelerates convergence by 37.2% on average and boosts final test accuracy by 2.1–5.8 percentage points over state-of-the-art federated learning baselines. The framework is particularly suited for latency-sensitive edge intelligence applications.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Search and Optimization: Learning to Search
• Humans and AI: Human-in-the-loop Machine Learning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

As privacy protection gains increasing importance, more models are being trained on edge devices and subsequently merged into the central server through Federated Learning (FL). However, current research overlooks the impact of network topology, physical distance, and data heterogeneity on edge devices, leading to issues such as increased latency and degraded model performance. To address these issues, we propose a new federated learning scheme on edge devices that called Federated Learning with Encrypted Data Sharing(FedEDS). FedEDS uses the client model and the model's stochastic layer to train the data encryptor. The data encryptor generates encrypted data and shares it with other clients. The client uses the corresponding client's stochastic layer and encrypted data to train and adjust the local model. FedEDS uses the client's local private data and encrypted shared data from other clients to train the model. This approach accelerates the convergence speed of federated learning training and mitigates the negative impact of data heterogeneity, making it suitable for application services deployed on edge devices requiring rapid convergence. Experiments results show the efficacy of FedEDS in promoting model performance.

Problem

• Addresses latency and performance issues in federated learning
• Mitigates data heterogeneity impact on edge devices
• Enhances convergence speed with encrypted data sharing

Innovation

• Encrypted data sharing among edge devices
• Client model trains data encryptor
• Combines private and shared encrypted data