FedPETv1: Edge-aware federated learning with differential privacy for distributed medical diagnostics

Li, Shancang, Zhao, Shanshan and Ding, Yongjian 2025. FedPETv1: Edge-aware federated learning with differential privacy for distributed medical diagnostics. Presented at: The 31st Annual International Conference on Mobile Computing and Networking, Hong Kong, China, 4-8 November 2025. Proceedings of the Federated Learning and Edge AI for Privacy and Mobility. ACM, pp. 4-7. 10.1145/3737899.3768515

Full text not available from this repository.

Abstract

In this work we propose FedPET, a privacy-enhancing federated learning (FL) framework designed for decentralized healthcare analytics. FedPET iteratively aggregates locally computed healthcare analytics model updates from edge devices and then conduct collaborative model training across distributed healthcare partners without sharing raw healthcare data. FedPET employs adaptive aggregation weights and momentum-based optimization to address non-IID data distributions while minimizing communication overhead. Experiment results demonstrated that FedPET achieves comparable accuracy to centralised training while reducing communication costs by 40%.

Item Type:	Conference or Workshop Item (Paper)
Date Type:	Publication
Status:	Published
Schools:	Schools > Computer Science & Informatics
Publisher:	ACM
ISBN:	9798400719769
Last Modified:	18 Dec 2025 10:30
URI:	https://orca.cardiff.ac.uk/id/eprint/183344

Actions (repository staff only)

Edit Item