《计算机应用》唯一官方网站 ›› 2025, Vol. 45 ›› Issue (12): 3747-3756.DOI: 10.11772/j.issn.1001-9081.2024121794

• 人工智能 •    下一篇

基于个性化子模型和K均值聚类的联邦学习公平性算法

景忠瑞1,2,3, 陈学斌1,2,3, 菅银龙1,2,3, 钟琪1,2,3, 张镇博1,2,3   

  1. 1.华北理工大学 理学院,河北 唐山 063210
    2.河北省数据科学与应用重点实验室(华北理工大学),河北 唐山 063210
    3.唐山市数据科学重点实验室(华北理工大学),河北 唐山 063210
  • 收稿日期:2024-12-20 修回日期:2025-03-13 接受日期:2025-03-18 发布日期:2025-03-27 出版日期:2025-12-10
  • 通讯作者: 陈学斌
  • 作者简介:景忠瑞(2000—),男,山西临汾人,硕士研究生,CCF会员,主要研究方向:数据安全、隐私保护
    陈学斌(1970—),男,河北唐山人,教授,博士,CCF杰出会员,主要研究方向:大数据安全、物联网安全、网络安全
    菅银龙(2001—),男,河南商丘人,硕士研究生,CCF会员,主要研究方向:数据安全、隐私保护
    钟琪(1999—),女,河北张家口人,硕士研究生,CCF会员,主要研究方向:数据安全、隐私保护
    张镇博(1999—),男,山东济南人,硕士研究生,CCF会员,主要研究方向:数据安全、隐私保护。
  • 基金资助:
    国家自然科学基金资助项目(U20A20179)

Federated learning fairness algorithm based on personalized submodel and K-means clustering

Zhongrui JING1,2,3, Xuebin CHEN1,2,3, Yinlong JIAN1,2,3, Qi ZHONG1,2,3, Zhenbo ZHANG1,2,3   

  1. 1.College of Science,North China University of Science and Technology,Tangshan Hebei 063210,China
    2.Hebei Provincial Key Laboratory of Data Science and Application (North China University of Science and Technology),Tangshan Hebei 063210,China
    3.Tangshan Key Laboratory of Data Science (North China University of Science and Technology),Tangshan Hebei 063210,China
  • Received:2024-12-20 Revised:2025-03-13 Accepted:2025-03-18 Online:2025-03-27 Published:2025-12-10
  • Contact: Xuebin CHEN
  • About author:JING Zhongrui, born in 2000, M. S. candidate. His research interests include data security, privacy protection.
    CHEN Xuebin, born in 1970, Ph. D., professor. His research interests include big data security, internet of things security, network security.
    JIAN Yinlong, born in 2001, M. S. candidate. His research interests include data security, privacy protection.
    ZHONG Qi, born in 1999, M. S. candidate. Her research interests include data security, privacy protection.
    ZHANG Zhenbo, born in 1999, M. S. candidate. His research interests include data security, privacy protection.
  • Supported by:
    National Natural Science Foundation of China(U20A20179)

摘要:

传统联邦学习(FL)未考虑协作公平性,导致客户端获得的奖励与它的实际贡献不匹配。针对这一问题,提出一种基于个性化子模型和K均值聚类的联邦学习公平性算法(FedPSK)。首先,根据神经网络中神经元的激活模式对神经元聚类,且仅对聚类后的簇中心神经元进行重要性评估,并使用簇中心神经元的评分代表簇中其他神经元的评分,从而降低神经元评估的耗时;其次,使用层次选取方式选择客户端子模型中包含的神经元数量及编号,并为每个客户端建立具有完整神经网络结构的子模型;最后,通过为客户端下发子模型,实现协作公平性。在不同数据集上的实验结果表明,在公平度量的相关系数方面,FedPSK比FedSAC(Federated learning framework with dynamic Submodel Allocation for Collaborative fairness)提高了2.70%;在时间开销方面,FedPSK比FedSAC至少降低了84.12%。可见,FedPSK在提升FL算法公平性的同时,极大地降低了算法运行的时间开销,验证了所提算法的高效性。

关键词: 联邦学习, 协作公平性, K均值聚类, 子模型下发, 个性化

Abstract:

Traditional Federated Learning (FL) does not consider collaborative fairness, leading to a mismatch between the reward obtained by the client and its actual contribution. To address this issue, a Federated learning fairness algorithm based on Personalized Submodel and K-means clustering (FedPSK) was proposed. Firstly, the neurons in the neural network were clustered according to their activation patterns, and only the importance of the cluster center neurons after clustering was evaluated. And the scores of the cluster center neurons were used to represent the scores of other neurons in the cluster, which reduced the time consumption of neuron evaluation. Then, the number of neurons and their labeling included in the client submodel were selected through hierarchical selection method, and a submodel with a complete neural network structure was constructed for each client. Finally, collaborative fairness was achieved by distributing submodels to the clients. Experimental results on different datasets show that FedPSK improves the correlation coefficient of fairness measurement by 2.70% compared with FedSAC (Federated learning framework with dynamic Submodel Allocation for Collaborative fairness). In terms of time overhead, FedPSK reduces by at least 84.12% compared with FedSAC. It can be seen that FedPSK improves the fairness of FL algorithm, and reduces the time overhead of the algorithm execution greatly, verifying the efficiency of the proposed algorithm.

Key words: Federated Learning (FL), collaborative fairness, K-means clustering, submodel distribution, personalization

中图分类号: