Personalized Decentralized Federated Learning with Knowledge Distillation

被引:2
作者
Jeong, Eunjeong [1 ]
Kountouris, Marios [1 ]
机构
[1] EURECOM, Commun Syst Dept, F-06410 Sophia Antipolis, France
来源
ICC 2023-IEEE INTERNATIONAL CONFERENCE ON COMMUNICATIONS | 2023年
关键词
decentralized federated learning; personalization; knowledge distillation;
D O I
10.1109/ICC45041.2023.10279714
中图分类号
TN [电子技术、通信技术];
学科分类号
0809 ;
摘要
Personalization in federated learning (FL) functions as a coordinator for clients with high variance in data or behavior. Ensuring the convergence of these clients' models relies on how closely users collaborate with those with similar patterns or preferences. However, it is generally challenging to quantify similarity under limited knowledge about other users' models given to users in a decentralized network. To cope with this issue, we propose a personalized and fully decentralized FL algorithm, leveraging knowledge distillation techniques to empower each device so as to discern statistical distances between local models. Each client device can enhance its performance without sharing local data by estimating the similarity between two intermediate outputs from feeding local samples as in knowledge distillation. Our empirical studies demonstrate that the proposed algorithm improves the test accuracy of clients in fewer iterations under highly non-independent and identically distributed (non-i.i.d.) data distributions and is beneficial to agents with small datasets, even without the need for a central server.
引用
收藏
页码:1982 / 1987
页数:6
相关论文
共 50 条
  • [31] FEDGKD: Toward Heterogeneous Federated Learning via Global Knowledge Distillation
    Yao, Dezhong
    Pan, Wanning
    Dai, Yutong
    Wan, Yao
    Ding, Xiaofeng
    Yu, Chen
    Jin, Hai
    Xu, Zheng
    Sun, Lichao
    IEEE TRANSACTIONS ON COMPUTERS, 2024, 73 (01) : 3 - 17
  • [32] Heterogeneous Federated Learning Framework for IIoT Based on Selective Knowledge Distillation
    Guo, Sheng
    Chen, Hui
    Liu, Yang
    Yang, Chengyi
    Li, Zengxiang
    Jin, Cheng Hao
    IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS, 2025, 21 (02) : 1078 - 1089
  • [33] Parameterized data-free knowledge distillation for heterogeneous federated learning
    Guo, Cheng
    He, Qianqian
    Tang, Xinyu
    Liu, Yining
    Jie, Yingmo
    KNOWLEDGE-BASED SYSTEMS, 2025, 317
  • [34] Prototype-Decomposed Knowledge Distillation for Learning Generalized Federated Representation
    Wu, Aming
    Yu, Jiaping
    Wang, Yuxuan
    Deng, Cheng
    IEEE TRANSACTIONS ON MULTIMEDIA, 2024, 26 : 10991 - 11002
  • [35] Communication-efficient Federated Learning for UAV Networks with Knowledge Distillation and Transfer Learning
    Li, Yalong
    Wu, Celimuge
    Du, Zhaoyang
    Zhong, Lei
    Yoshinaga, Tsutomu
    IEEE CONFERENCE ON GLOBAL COMMUNICATIONS, GLOBECOM, 2023, : 5739 - 5744
  • [36] Sparse Personalized Federated Learning
    Liu, Xiaofeng
    Li, Yinchuan
    Wang, Qing
    Zhang, Xu
    Shao, Yunfeng
    Geng, Yanhui
    IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2024, 35 (09) : 12027 - 12041
  • [37] Decentralized and Model-Free Federated Learning: Consensus-Based Distillation in Function Space
    Taya, Akihito
    Nishio, Takayuki
    Morikura, Masahiro
    Yamamoto, Koji
    IEEE TRANSACTIONS ON SIGNAL AND INFORMATION PROCESSING OVER NETWORKS, 2022, 8 : 799 - 814
  • [38] Incentive and Knowledge Distillation Based Federated Learning for Cross-Silo Applications
    Li, Beibei
    Shi, Yaxin
    Guo, Yuqing
    Kong, Qinglei
    Jiang, Yukun
    IEEE INFOCOM 2022 - IEEE CONFERENCE ON COMPUTER COMMUNICATIONS WORKSHOPS (INFOCOM WKSHPS), 2022,
  • [39] One Teacher is Enough: A Server-Clueless Federated Learning With Knowledge Distillation
    Ning, Wanyi
    Qi, Qi
    Wang, Jingyu
    Zhu, Mengde
    Li, Shaolong
    Yang, Guang
    Liao, Jianxin
    IEEE TRANSACTIONS ON SERVICES COMPUTING, 2024, 17 (05) : 2704 - 2718
  • [40] FedAL: Black-Box Federated Knowledge Distillation Enabled by Adversarial Learning
    Han, Pengchao
    Shi, Xingyan
    Huang, Jianwei
    IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, 2024, 42 (11) : 3064 - 3077