Trading Off Privacy, Utility and Efficiency in Federated Learning

TL;DR

This paper introduces a unified framework for federated learning that models the inherent trade-offs among privacy, utility, and efficiency, providing theoretical bounds and guidance for protection mechanisms.

Contribution

It formulates a comprehensive trade-off analysis in federated learning, establishing the NFL theorem and analyzing lower bounds for key protection mechanisms.

Findings

Trade-offs between privacy, utility, and efficiency are quantified.

NFL theorem states simultaneous optimization of all three is unrealistic.

Lower bounds for protection mechanisms guide parameter selection.

Abstract

Federated learning (FL) enables participating parties to collaboratively build a global model with boosted utility without disclosing private data information. Appropriate protection mechanisms have to be adopted to fulfill the opposing requirements in preserving \textit{privacy} and maintaining high model \textit{utility}. In addition, it is a mandate for a federated learning system to achieve high \textit{efficiency} in order to enable large-scale model training and deployment. We propose a unified federated learning framework that reconciles horizontal and vertical federated learning. Based on this framework, we formulate and quantify the trade-offs between privacy leakage, utility loss, and efficiency reduction, which leads us to the No-Free-Lunch (NFL) theorem for the federated learning system. NFL indicates that it is unrealistic to expect an FL algorithm to simultaneously provide excellent privacy, utility, and efficiency in certain scenarios. We then analyze the lower bounds for the privacy leakage, utility loss and efficiency reduction for several widely-adopted protection mechanisms including \textit{Randomization}, \textit{Homomorphic Encryption}, \textit{Secret Sharing} and \textit{Compression}. Our analysis could serve as a guide for selecting protection parameters to meet particular requirements.