文献类型：期刊文献

英文题名：Do Larger Batch Sizes Always Help? Revisiting Incentive-Driven Differential Privacy Federated Learning

作者：Xu, Jin[1];Yu, Huiqun[1,2];Fan, Guisheng[1,2];Zhang, Hengrun[1]

机构：[1]East China Univ Sci & Technol, Dept Comp Sci & Engn, Shanghai 200237, Peoples R China;[2]Shanghai Engn Res Ctr Smart Energy, Shanghai 201209, Peoples R China

年份：2026

卷号：12

起止页码：1862

外文期刊名：IEEE TRANSACTIONS ON COGNITIVE COMMUNICATIONS AND NETWORKING

收录：;EI(收录号：20253519055606);WOS:【SCI-EXPANDED(收录号:WOS:001652009800043)】；

基金：This work was partially supported by the NationalNatural Science Foundation of China (No. 62372174), the Shanghai Sci-ence and Technology Commission Computational Biology Program (No.23JS1400600), the National Engineering Laboratory Research Program forBig Data Distribution and Exchange Technologies (No. 2021-GYHLW-01007), the Shanghai 2024 Science and Technology Innovation ActionPlan Star Cultivation (Sailing Program, No. 24YF2720000), the Sci-ence Foundation of Shanghai Key Laboratory of Computer SoftwareEvaluating and Testing (No. SSTL2024_02), and the Science Founda-tion of State Key Laboratory for Novel Software Technology, Nan-jing University, P.R. China (No. KFKT2025B82).

语种：英文

外文关键词：Privacy; Convergence; Training; Noise; Federated learning; Differential privacy; Servers; Optimization; Data models; Computational modeling; differential privacy; client selection; batch size optimization; Stackelberg game

摘要：Differential Privacy Federated Learning (DP-FL) combines Differential Privacy (DP) with Federated Learning (FL), enabling multiple clients to collaboratively train a shared model while protecting data privacy. However, introducing DP into FL will add noise to model parameters, which typically deteriorates model convergence. Although a recent work has revealed the compensation effect by increasing the total batch size, it overlooks the "generalization gap" phenomenon, which is induced by excessively large batch size and has been long discussed in the machine learning field. In order to avoid the other extreme, we strengthen several core components in, and propose an Incentive-driven Differential Privacy Federated Learning (IDP-FL) framework. First, instead of building all hopes on batch sizes, the proposed framework jointly considers the non-IID degrees of local data and clients' privacy budgets, minimizing the difference between the optimal batch size for each selected client and its corresponding critical batch size. Second, we reconfigure the batch size for each selected client by balancing the negative impact of DP noise on convergence and of the "generalization gap" phenomenon. Finally, we design a Stackelberg game-based incentive mechanism that encourages clients to contribute computational resources, and prove the existence of a Stackelberg equilibrium to guarantee stability. Through numerical evaluations on real-world datasets, we show that our IDP-FL framework outperforms existing algorithms in terms of test accuracy and utility. Ablation studies further confirm the effectiveness of each component.