Newest papers

Privacy-Preserving Personalized Federated Learning (PPP-FL) for Medical Imaging

Md Morshedul Islam, Hossain Shahriar, Alfredo Cuzzocrea · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.27470 · cited by 0 works (OpenAlex) · open-access PDF

French

L’apprentissage à portée fédérée (FL) est un paradigme d’apprentissage par machine distribué qui réduit les coûts de transmission, atténue les restrictions de partage de données transfrontalières et améliore la confidentialité en conservant les données brutes sur les appareils locaux. Les clients dans les systèmes FL développent des modèles locaux et ne partagent que des mises à jour, qui sont agrégées par un serveur central. Alors que cette configuration limite l’exposition directe aux données, FL reste vulnérable aux attaques de confidentialité. Les stratégies d’attaque et de défense existantes ont été largement développées pour les systèmes centralisés et leur adaptation à FL introduit de nouveaux défis en raison de son architecture unique et de son flux de travail

Machine translation (Meta M2M100); the English original follows.

Abstract

Federated learning (FL) is a distributed machine learning paradigm that reduces transmission costs, mitigates cross-border data sharing restrictions, and enhances privacy by keeping raw data on local devices. Clients in FL systems train local models and share only updates, which are aggregated by a central server. While this setup limits direct data exposure, FL remains vulnerable to privacy attacks. Existing attack and defense strategies have largely been developed for centralized systems, and adapting them to FL introduces new challenges due to its unique architecture and training workflow. In this work, we first propose a model inversion attack on FL systems using Deep Convolutional Generative Adversarial Networks (DCGANs) in the context of large-scale medical imaging. Our method operates under realistic assumptions about the attacker's knowledge and capabilities. We then examine the effect of client-side differential privacy (DP) in mitigating privacy leakage and incorporate personalized FL to improve model performance under DP settings. Experiments on the Brain MRI ND-5 dataset show that DCGAN-based attacks can reconstruct semantically meaningful synthetic images, revealing significant risks to the privacy of training data. Applying DP reduces attack success but introduces an accuracy trade-off, which is further overcome by personalized FL at each client side. We also evaluate how different aggregation algorithms affect both attack vulnerability and defense effectiveness, underscoring the critical role of aggregation strategies. The proposed attack-and-defense framework is model-agnostic and applicable to a wide range of FL-based systems.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Federated learning, Differential privacy, Adversarial system, Context (archaeology), Raw data

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange