Privacy-Preserving Federated Primal—Dual Learning for Nonconvex and Nonsmooth Problems With Model Sparsification

Federated learning (FL) has been recognized as a rapidly growing research area, where the model is trained over massively distributed clients under the orchestration of a parameter server (PS) without sharing clients’ data. This article delves into a class of federated problems characterized by nonconvex and nonsmooth (NCNS) loss functions, that are prevalent in FL applications but challenging to handle due to their intricate nonconvexity and nonsmoothness nature and the conflicting requirements on communication efficiency and privacy protection. In this article, we propose a novel federated primal—dual algorithm with bidirectional model sparsification tailored for NCNS FL problems, and differential privacy is applied for privacy guarantee. Its unique insightful properties and some privacy and convergence analyses are also presented as the FL algorithm design guidelines. Extensive experiments on real-world data are conducted to demonstrate the effectiveness of the proposed algorithm and much superior performance than some state-of-the-art FL algorithms, together with the validation of all the analytical results and properties.