JSAM: Privacy Straggler-Resilient Joint Client Selection and Incentive Mechanism Design in Differentially Private Federated Learning

TL;DR

This paper introduces JSAM, a Bayesian framework for optimizing client selection and privacy compensation in differentially private federated learning, improving accuracy and efficiency by selectively including privacy-tolerant clients.

Contribution

JSAM is the first to jointly optimize client selection and privacy compensation in federated learning, transforming complex problems into an efficient three-dimensional solution.

Findings

JSAM improves test accuracy by up to 15% over existing methods.

Selective client inclusion enhances training efficiency and cost-effectiveness.

Clients with minimal privacy sensitivity may face higher cumulative costs.

Abstract

Differentially private federated learning faces a fundamental tension: privacy protection mechanisms that safeguard client data simultaneously create quantifiable privacy costs that discourage participation, undermining the collaborative training process. Existing incentive mechanisms rely on unbiased client selection, forcing servers to compensate even the most privacy-sensitive clients ("privacy stragglers"), leading to systemic inefficiency and suboptimal resource allocation. We introduce JSAM (Joint client Selection and privacy compensAtion Mechanism), a Bayesian-optimal framework that simultaneously optimizes client selection probabilities and privacy compensation to maximize training effectiveness under budget constraints. Our approach transforms a complex 2N-dimensional optimization problem into an efficient three-dimensional formulation through novel theoretical characterization of optimal selection strategies. We prove that servers should preferentially select privacy-tolerant clients while excluding high-sensitivity participants, and uncover the counter-intuitive insight that clients with minimal privacy sensitivity may incur the highest cumulative costs due to frequent participation. Extensive evaluations on MNIST and CIFAR-10 demonstrate that JSAM achieves up to 15% improvement in test accuracy compared to existing unbiased selection mechanisms while maintaining cost efficiency across varying data heterogeneity levels.