文献类型：期刊文献

英文题名：VDSV: Client Selection in Federated Learning Based on Value Density and Secondary Verification

作者：Ding, Weichao[1];Zhou, Zhou[1];Min, Qi[1];Luo, Fei[1];Dong, Wenbo[1];Zhang, Hengrun[1]

机构：[1]East China Univ Sci & Technol, Sch Informat Sci & Engn, Shanghai 200231, Peoples R China

年份：2026

卷号：23

起止页码：699

外文期刊名：IEEE TRANSACTIONS ON NETWORK AND SERVICE MANAGEMENT

收录：;EI(收录号：20254819622354);WOS:【SCI-EXPANDED(收录号:WOS:001655693000015)】；

基金：This work is sponsored by the National Natural Science Foundation of China (No. 62403201); the Shanghai Pilot Program for Basic Research (22TQ1400100-16); the National Key Research and Development Program of China (2024YFC3307700); Nature Science Foundation of Shanghai (24ZR1415200, 23ZR1414900, 22ZR1416500); and the Science Foundation of State Key Laboratory for Novel Software Technology, Nanjing University, P. R. China under Grant KFKT2025B82.

语种：英文

外文关键词：Convergence; Training; Data models; Servers; Distributed databases; Analytical models; Interference; Federated learning; Costs; Artificial intelligence; client selection; data heterogeneity; value density; secondary verification

摘要：Client selection has been widely considered in Federated Learning (FL) to reduce communication overhead while ensuring proper convergence performance. Due to data heterogeneity in FL, a representative subset of participants should take into account both intra- and inter-client diversity. While existing works usually emphasize on one of them, this paper proposes a VDSV (client selection based on Value Density and Secondary Verification) framework, which optimizes the client selection strategy from both sides. Therein, intra- and inter-client diversity are respectively measured based on a designed client data score as well as gradient distance and direction. Afterwards, a client selection model is established based on a proposed metric, called client value density. Besides, a secondary validation method is developed to dynamically tweak the current client selection and model aggregation strategies. The general idea of the above design is based on the theoretical convergence analysis and the observation that the client contribution to the global model can get changed throughout the learning process. The experimental results demonstrate that VDSV can achieve higher convergence rates and ensure comparable model performance. In specific, our method can reduce the communication rounds by an average of 37.88%, which saves noticeable communication overhead.