Breaking the Communication-Privacy-Accuracy Tradeoff with $f$-Differential Privacy

TL;DR

This paper introduces a novel approach using $f$-differential privacy to enhance privacy guarantees in federated data analysis, effectively balancing privacy, communication efficiency, and accuracy.

Contribution

It derives tight $f$-DP guarantees for discrete mechanisms, proposes a ternary compressor for privacy amplification, and breaks the traditional tradeoff between privacy, communication, and accuracy.

Findings

Tight $f$-DP bounds for binomial and sign-based mechanisms.

Improved distributed mean estimation without accuracy loss.

Enhanced privacy amplification via sparsification.

Abstract

We consider a federated data analytics problem in which a server coordinates the collaborative data analysis of multiple users with privacy concerns and limited communication capability. The commonly adopted compression schemes introduce information loss into local data while improving communication efficiency, and it remains an open problem whether such discrete-valued mechanisms provide any privacy protection. In this paper, we study the local differential privacy guarantees of discrete-valued mechanisms with finite output space through the lens of $f$-differential privacy (DP). More specifically, we advance the existing literature by deriving tight $f$-DP guarantees for a variety of discrete-valued mechanisms, including the binomial noise and the binomial mechanisms that are proposed for privacy preservation, and the sign-based methods that are proposed for data compression, in closed-form expressions. We further investigate the amplification in privacy by sparsification and propose a ternary stochastic compressor. By leveraging compression for privacy amplification, we improve the existing methods by removing the dependency of accuracy (in terms of mean square error) on communication cost in the popular use case of distributed mean estimation, therefore breaking the three-way tradeoff between privacy, communication, and accuracy. Finally, we discuss the Byzantine resilience of the proposed mechanism and its application in federated learning.