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

Spanish

El aprendizaje de rango federado (FL) es un paradigma de aprendizaje de marcos de datos distribuido que reduce los costos de transmisión, mitiga las restricciones de intercambio de datos transfronterizos y mejora la privacidad al mantener los datos brutos en los dispositivos locales. Los clientes en los sistemas FL entrenan los modelos locales y comparten solo las actualizaciones, que son agregadas por un servidor central. Mientras que esta configuración limita la exposición directa a los datos, FL sigue siendo vulnerable a los ataques de privacidad. Las estrategias de ataque y defensa existentes se han desarrollado en gran medida para los sistemas centralizados, y adaptarlos a los sistemas FL introduce nuevos desafíos debido a su arquitectura y flujo de trabajo agregado único. En este trabajo, primero proponemos un ataque de modelo de inversión a los sistemas FL utilizando Redes Adversarias Gener

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