Federated $f$-Differential Privacy
This work addresses privacy concerns for clients in federated learning systems, offering a novel privacy notion and framework, though it appears incremental as it builds on existing differential privacy concepts.
The paper tackles the problem of enhancing privacy in federated learning by introducing federated f-differential privacy, a new notion tailored to the federated setting, and proposes a framework called PriFedSync that provably achieves this privacy guarantee while empirically demonstrating trade-offs with prediction performance in computer vision tasks.
Federated learning (FL) is a training paradigm where the clients collaboratively learn models by repeatedly sharing information without compromising much on the privacy of their local sensitive data. In this paper, we introduce federated $f$-differential privacy, a new notion specifically tailored to the federated setting, based on the framework of Gaussian differential privacy. Federated $f$-differential privacy operates on record level: it provides the privacy guarantee on each individual record of one client's data against adversaries. We then propose a generic private federated learning framework {PriFedSync} that accommodates a large family of state-of-the-art FL algorithms, which provably achieves federated $f$-differential privacy. Finally, we empirically demonstrate the trade-off between privacy guarantee and prediction performance for models trained by {PriFedSync} in computer vision tasks.