Client Selection With Staleness Compensation in Asynchronous Federated Learning

被引:4
|
作者
Zhu, Hongbin [1 ]
Kuang, Junqian [2 ]
Yang, Miao [3 ]
Qian, Hua [2 ]
机构
[1] Fudan Univ, Inst FinTech, Shanghai 200082, Peoples R China
[2] Chinese Acad Sci, Shanghai Adv Res Inst, Shanghai 201210, Peoples R China
[3] ShanghaiTech Univ, Sch Informat Sci & Technol SIST, Shanghai 201210, Peoples R China
基金
中国国家自然科学基金;
关键词
Asynchronous federated learning; staleness compensation; client selection; multi-armed bandit;
D O I
10.1109/TVT.2022.3220809
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
As a nascent privacy-preserving machine learning (ML) paradigm, federated learning (FL) leverages distributed clients at the network edge to collaboratively train an ML model. Asynchronous FL overcomes the straggler issue in synchronous FL. However, asynchronous FL incurs the staleness problem, which degrades the training performance of FL over wireless networks. To tackle the staleness problem, we develop a staleness compensation algorithm to improve the training performance of FL in terms of convergence and test accuracy. By including the first-order term in Taylor expansion of the gradient function, the proposed algorithm compensates the staleness in asynchronous FL. To further minimize training latency, we model the client selection for asynchronous FL as a multi-armed bandit problem. We develop an online client selection algorithm to minimize training latency without prior knowledge of the channel condition or local computing status. Simulation results show that the proposed algorithm outperforms the baseline algorithms in both test accuracy and training latency.
引用
收藏
页码:4124 / 4129
页数:6
相关论文
共 50 条
  • [41] Adaptive client selection with personalization for communication efficient Federated Learning
    de Souza, Allan M.
    Maciel, Filipe
    da Costa, Joahannes B. D.
    Bittencourt, Luiz F.
    Cerqueira, Eduardo
    Loureiro, Antonio A. F.
    Villas, Leandro A.
    [J]. AD HOC NETWORKS, 2024, 157
  • [42] Joint Client Selection and Bandwidth Allocation Algorithm for Federated Learning
    Ko, Haneul
    Lee, Jaewook
    Seo, Sangwon
    Pack, Sangheon
    Leung, Victor C. M.
    [J]. IEEE TRANSACTIONS ON MOBILE COMPUTING, 2023, 22 (06) : 3380 - 3390
  • [43] Auction-based client selection for online Federated Learning
    Guo, Juncai
    Su, Lina
    Liu, Jin
    Ding, Jianli
    Liu, Xiao
    Huang, Bo
    Li, Li
    [J]. INFORMATION FUSION, 2024, 112
  • [44] FedECS: Client Selection for Optimizing Computing Energy in Federated Learning
    Han, Shuo
    Zhang, Chenyu
    Wang, Luhan
    Zheng, Wei
    Wen, Xiangming
    [J]. 2023 IEEE 34TH ANNUAL INTERNATIONAL SYMPOSIUM ON PERSONAL, INDOOR AND MOBILE RADIO COMMUNICATIONS, PIMRC, 2023,
  • [45] Adaptive client selection and model aggregation for heterogeneous federated learning
    Zhai, Rui
    Jin, Haozhe
    Gong, Wei
    Lu, Ke
    Liu, Yanhong
    Song, Yalin
    Yu, Junyang
    [J]. MULTIMEDIA SYSTEMS, 2024, 30 (04)
  • [46] Client Selection Based on Label Quantity Information for Federated Learning
    Ma, Jiahua
    Sun, Xinghua
    Xia, Wenchao
    Wang, Xijun
    Chen, Xiang
    Zhu, Hongbo
    [J]. 2021 IEEE 32ND ANNUAL INTERNATIONAL SYMPOSIUM ON PERSONAL, INDOOR AND MOBILE RADIO COMMUNICATIONS (PIMRC), 2021,
  • [47] Towards Instant Clustering Approach for Federated Learning Client Selection
    Arisdakessian, Sarhad
    Wahab, Omar Abdel
    Mourad, Azzam
    Otrok, Hadi
    [J]. 2023 INTERNATIONAL CONFERENCE ON COMPUTING, NETWORKING AND COMMUNICATIONS, ICNC, 2023, : 409 - 413
  • [48] Federated Learning with Personalized Differential Privacy Combining Client Selection
    Xie, Yunting
    Zhang, Lan
    [J]. 2022 8TH INTERNATIONAL CONFERENCE ON BIG DATA COMPUTING AND COMMUNICATIONS, BIGCOM, 2022, : 79 - 87
  • [49] Blockchain-based Secure Client Selection in Federated Learning
    Nguyen, Truc
    Thai, Phuc
    Jeter, Tre R.
    Dinht, Thang N.
    Thai, My T.
    [J]. 2022 IEEE INTERNATIONAL CONFERENCE ON BLOCKCHAIN AND CRYPTOCURRENCY (IEEE ICBC 2022), 2022,
  • [50] Welfare and fairness dynamics in federated learning: a client selection perspective
    Travadi, Yash
    Peng, Le
    Bi, Xuan
    Sun, Ju
    Yang, Mochen
    [J]. STATISTICS AND ITS INTERFACE, 2024, 17 (03) : 383 - 395