Share Your Representation Only: Guaranteed Improvement of the Privacy-Utility Tradeoff in Federated Learning

TL;DR

This paper introduces a new federated learning approach that guarantees privacy while improving the utility of learned representations, achieving better privacy-utility trade-offs through a novel algorithm and theoretical analysis.

Contribution

Proposes \\DPFEDREP, a differentially private federated learning algorithm with convergence guarantees and improved utility-privacy trade-off in the linear representation setting.

Findings

Converges to a neighborhood of the global optimum at a linear rate.

Improves the state-of-the-art utility-privacy trade-off by a factor of \\sqrt{d}.

Demonstrates significant empirical performance gains on image classification tasks.

Abstract

Repeated parameter sharing in federated learning causes significant information leakage about private data, thus defeating its main purpose: data privacy. Mitigating the risk of this information leakage, using state of the art differentially private algorithms, also does not come for free. Randomized mechanisms can prevent convergence of models on learning even the useful representation functions, especially if there is more disagreement between local models on the classification functions (due to data heterogeneity). In this paper, we consider a representation federated learning objective that encourages various parties to collaboratively refine the consensus part of the model, with differential privacy guarantees, while separately allowing sufficient freedom for local personalization (without releasing it). We prove that in the linear representation setting, while the objective is non-convex, our proposed new algorithm \DPFEDREP\ converges to a ball centered around the \emph{global optimal} solution at a linear rate, and the radius of the ball is proportional to the reciprocal of the privacy budget. With this novel utility analysis, we improve the SOTA utility-privacy trade-off for this problem by a factor of $\sqrt{d}$, where $d$ is the input dimension. We empirically evaluate our method with the image classification task on CIFAR10, CIFAR100, and EMNIST, and observe a significant performance improvement over the prior work under the same small privacy budget. The code can be found in this link: https://github.com/shenzebang/CENTAUR-Privacy-Federated-Representation-Learning.