网络与信息安全学报2025，Vol.11Issue(3)：175-189,15.DOI:10.11959/j.issn.2096-109x.2025029
面向异构数据的半分布式联邦学习安全聚合
Secure aggregation for semi-decentralized federated learning under heterogeneous data
摘要
Abstract
Semi-decentralized federated learning has gained attention for combining the advantages of centralized and decentralized approaches,thereby enhancing system scalability and flexibility.However,in the open heteroge-neous federated learning architecture,the existence of semi-honest participants and heterogeneous data has aggra-vated the problems of node privacy leakage and difficult model convergence during model aggregation.The differ-ences between local models have increased with the degree of local data heterogeneity,which has further slowed down the global model'convergence and reduced its accuracy.Additionally,the complex network topology among nodes has heightened the risk of local data privacy leakage,making it difficult to ensure local data privacy protec-tion.To tackle these issues,a secure aggregation scheme for heterogeneous data under a semi-distributed federated learning architecture was designed.This scheme aimed to enhance the global model's convergence performance while safeguarding node privacy.A random masking mechanism was developed based on the alternating direction multiplier method.This mechanism strengthened local data privacy protection and prevented access to individual node models.Moreover,a double-weighted aggregation strategy was proposed.In this strategy,intra-cluster aggre-gation weights for nodes were determined based on the loss of the global model on different node samples,and global aggregation weights for different clusters were established according to their contribution to the global model.Extensive experiments were carried out on three public standard datasets.The results demonstrate that,com-pared with advanced schemes,the proposed method improves model convergence speed and accuracy in the con-text of heterogeneous data.关键词
半分布式联邦学习/隐私保护/加权聚合/数据异构Key words
semi-decentralized federated learning/privacy protection/weighted aggregation/heterogeneous data分类
计算机与自动化引用本文复制引用
黄梅,王玲玲,张政胤,刘宇飞,孙艺桐..面向异构数据的半分布式联邦学习安全聚合[J].网络与信息安全学报,2025,11(3):175-189,15.基金项目
国家自然科学基金(61802217) （61802217）
山东省自然科学基金(ZR2023MF082) （ZR2023MF082）
青岛科技计划重点研发项目(22-3-4-xxgg-10-gx) The National Natural Science Foundation of China(61802217) （22-3-4-xxgg-10-gx）
The Natural Science Foundation of Shan-dong Province(ZR2023MF082) （ZR2023MF082）
Qingdao Science and Technology Program Key Research and Development Project(22-3-4-xxgg-10-gx) （22-3-4-xxgg-10-gx）
