Federated Learning with Enhanced Privacy Guarantees in Healthcare Systems
Elias Chias (Lead), Sarah Lin, Michael Chen
Published: March 2025
Journal of Machine Learning Research
Open Access
Abstract
This paper presents Federated Learning with Enhanced Privacy Guarantees (F-EPG), a novel framework for decentralized machine learning in healthcare systems. By integrating differential privacy with secure multi-party computation, our approach achieves a 47% improvement in model accuracy while maintaining privacy guarantees across 28 hospital networks. Experimental results show F-EPG outperforms existing methods in both data utility and privacy preservation.
Key Contributions
- • Novel privacy-preserving aggregation mechanism
- • Decentralized model training with minimal trust assumptions
- • Real-world deployments across 28 hospital systems
Impact
- • Citations: 214
- • Research Interest: 1.2k
- • Saved by: 483 researchers
Related Publications
Privacy-Preserving Federated Learning
Elias Chias et al. 2024
Differential Privacy in Health AI
Sarah Lin et al. 2024
Secure Multi-Party Computation
Michael Chen et al. 2023
Citation
@article{chias2025federated,
title={Federated Learning with Enhanced Privacy Guarantees in Healthcare Systems},
author={Chias, Elias and Lin, Sarah and Chen, Michael},
journal={Journal of Machine Learning Research},
year={2025},
volume={26},
pages={5103--5148},
publisher={Microtome}
}