Privacy Amplification via Random Participation in Federated Learning

TL;DR

This paper analyzes how random client participation in federated learning enhances differential privacy, especially when local datasets are small, and compares it to local subsampling privacy guarantees.

Contribution

It provides a theoretical analysis of privacy amplification through random participation in federated learning, considering non-uniform subsampling and dataset size effects.

Findings

Random participation improves privacy guarantees over local subsampling.

Privacy amplification is close to centralized setting for small datasets.

Large datasets may risk client identity disclosure despite random participation.

Abstract

Running a randomized algorithm on a subsampled dataset instead of the entire dataset amplifies differential privacy guarantees. In this work, in a federated setting, we consider random participation of the clients in addition to subsampling their local datasets. Since such random participation of the clients creates correlation among the samples of the same client in their subsampling, we analyze the corresponding privacy amplification via non-uniform subsampling. We show that when the size of the local datasets is small, the privacy guarantees via random participation is close to those of the centralized setting, in which the entire dataset is located in a single host and subsampled. On the other hand, when the local datasets are large, observing the output of the algorithm may disclose the identities of the sampled clients with high confidence. Our analysis reveals that, even in this case, privacy guarantees via random participation outperform those via only local subsampling.