Communication-Efficient Consensus Mechanism for Federated Reinforcement Learning

被引:3
|
作者
Xu, Xing [1 ]
Li, Rongpeng [1 ]
Zhao, Zhifeng [2 ]
Zhang, Honggang [1 ]
机构
[1] Zhejiang Univ, Hangzhou, Peoples R China
[2] Zhejiang Lab, Hangzhou, Peoples R China
基金
中国国家自然科学基金;
关键词
Independent Reinforcement Learning; Federated Learning; Consensus Algorithm; Communication Overheads; Utility Function;
D O I
10.1109/ICC45855.2022.9838936
中图分类号
TN [电子技术、通信技术];
学科分类号
0809 ;
摘要
The paper considers independent reinforcement learning (IRL) for multi-agent decision-making process in the paradigm of federated learning (FL). We show that FL can clearly improve the policy performance of IRL in terms of training efficiency and stability. However, since the policy parameters are trained locally and aggregated iteratively through a central server in FL, frequent information exchange incurs a large amount of communication overheads. To reach a good balance between improving the model's convergence performance and reducing the required communication and computation overheads, this paper proposes a system utility function and develops a consensus-based optimization scheme on top of the periodic averaging method, which introduces the consensus algorithm into FL for the exchange of a model's local gradients. This paper also provides novel convergence guarantees for the developed method, and demonstrates its superior effectiveness and efficiency in improving the system utility value through theoretical analyses and numerical simulation results.
引用
收藏
页码:80 / 85
页数:6
相关论文
共 50 条
  • [41] Dynamic Sampling and Selective Masking for Communication-Efficient Federated Learning
    Ji, Shaoxiong
    Jiang, Wenqi
    Walid, Anwar
    Li, Xue
    IEEE INTELLIGENT SYSTEMS, 2022, 37 (02) : 27 - 34
  • [42] SPinS-FL: Communication-Efficient Federated Subnetwork Learning
    Tsutsui, Masayoshi
    Takamaeda-Yamazaki, Shinya
    2023 IEEE 20TH CONSUMER COMMUNICATIONS & NETWORKING CONFERENCE, CCNC, 2023,
  • [43] Communication-efficient clustered federated learning via model distance
    Zhang, Mao
    Zhang, Tie
    Cheng, Yifei
    Bao, Changcun
    Cao, Haoyu
    Jiang, Deqiang
    Xu, Linli
    MACHINE LEARNING, 2024, 113 (06) : 3869 - 3888
  • [44] Time-Correlated Sparsification for Communication-Efficient Federated Learning
    Ozfatura, Emre
    Ozfatura, Kerem
    Gunduz, Deniz
    2021 IEEE INTERNATIONAL SYMPOSIUM ON INFORMATION THEORY (ISIT), 2021, : 461 - 466
  • [45] Adaptive Differential Filters for Fast and Communication-Efficient Federated Learning
    Becking, Daniel
    Kirchhoffer, Heiner
    Tech, Gerhard
    Haase, Paul
    Mueller, Karsten
    Schwarz, Heiko
    Samek, Wojciech
    2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION WORKSHOPS, CVPRW 2022, 2022, : 3366 - 3375
  • [46] LaF: Lattice-Based and Communication-Efficient Federated Learning
    Xu, Peng
    Hu, Manqing
    Chen, Tianyang
    Wang, Wei
    Jin, Hai
    IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, 2022, 17 : 2483 - 2496
  • [47] Selective Updates and Adaptive Masking for Communication-Efficient Federated Learning
    Herzog, Alexander
    Southam, Robbie
    Belarbi, Othmane
    Anwar, Saif
    Bullo, Marcello
    Carnelli, Pietro
    Khan, Aftab
    IEEE TRANSACTIONS ON GREEN COMMUNICATIONS AND NETWORKING, 2024, 8 (02): : 852 - 864
  • [48] VeriFL: Communication-Efficient and Fast Verifiable Aggregation for Federated Learning
    Guo, Xiaojie
    Liu, Zheli
    Li, Jin
    Gao, Jiqiang
    Hou, Boyu
    Dong, Changyu
    Baker, Thar
    IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, 2021, 16 : 1736 - 1751
  • [49] Communication-Efficient Federated Learning for Network Traffic Anomaly Detection
    Cui, Xiao
    Han, Xiaohui
    Liu, Guangqi
    Zuo, Wenbo
    Wang, Zhiwen
    2023 19TH INTERNATIONAL CONFERENCE ON MOBILITY, SENSING AND NETWORKING, MSN 2023, 2023, : 398 - 405
  • [50] Communication-Efficient Federated Learning Algorithm Based on Event Triggering
    Gao H.
    Yang L.
    Zhu J.
    Zhang M.
    Wu Q.
    Dianzi Yu Xinxi Xuebao/Journal of Electronics and Information Technology, 2023, 45 (10): : 3710 - 3718