Learn to Collaborate in MEC: An Adaptive Decentralized Federated Learning Framework

被引:2
作者
Wang, Yatong [1 ]
Wen, Zhongyi [2 ]
Li, Yunjie [1 ]
Cao, Bin [3 ]
机构
[1] Beijing Inst Technol, Sch Informat & Elect, Beijing 100081, Peoples R China
[2] Univ Elect Sci & Technol China, Sch Informat & Commun Engn, Chengdu 611731, Peoples R China
[3] Beijing Univ Posts & Telecommun, State Key Lab Networking & Switching Technol, Beijing 100876, Peoples R China
基金
中国国家自然科学基金; 中国博士后科学基金;
关键词
Collaboration; Training; Computational modeling; Computer architecture; Federated learning; Servers; Distributed databases; mobile edge computing; option-critic; multi-agent reinforcement learning;
D O I
10.1109/TMC.2024.3439588
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Decentralized federated learning (DFL) has emerged as a conducive paradigm, facilitating a distributed privacy-preserving data collaboration mode in mobile edge computing (MEC) systems to bolster the expansion of artificial intelligence applications. Nevertheless, the dynamic wireless environment and the heterogeneity among collaborating nodes, characterized by skewed datasets and uneven capabilities, present substantial challenges for efficient DFL model training in MEC systems. Consequently, the design of an efficient collaboration strategy becomes essential to facilitate practical distributed knowledge sharing and cost reduction for MEC. In this paper, we propose an adaptive decentralized federated learning framework that enables heterogeneous nodes to learn tailored collaboration strategies, thereby maximizing the efficiency of the DFL training process in collaborative MEC systems. Specifically, we present an effective option critic-based collaboration strategy learning (OCSL) mechanism by decomposing the collaboration strategy model into two sub-strategies: local training strategy and resource scheduling strategy. In addressing inherent issues such as large-scale action space and overestimation in collaboration strategy learning, we introduce the option framework and a dual critic network-based approximation method within the OCSL design. We theoretically prove that the learned collaboration strategy achieves the Nash equilibrium. Extensive numerical results demonstrate the effectiveness of the proposed method in comparison with existing baselines.
引用
收藏
页码:14071 / 14084
页数:14
相关论文
共 50 条
[41]   Enhancing Decentralized and Personalized Federated Learning With Topology Construction [J].
Chen, Suo ;
Xu, Yang ;
Xu, Hongli ;
Ma, Zhenguo ;
Wang, Zhiyuan .
IEEE TRANSACTIONS ON MOBILE COMPUTING, 2024, 23 (10) :9692-9707
[42]   Voting Consensus-Based Decentralized Federated Learning [J].
Gou, Yan ;
Weng, Shangyin ;
Imran, Muhammad Ali ;
Zhang, Lei .
IEEE INTERNET OF THINGS JOURNAL, 2024, 11 (09) :16267-16278
[43]   Decentralized Federated Learning Over Random Access Channel [J].
Kang, Yunseok ;
Song, Jaeyoung .
IEEE WIRELESS COMMUNICATIONS LETTERS, 2024, 13 (11) :3207-3211
[44]   A Novel Framework for the Analysis and Design of Heterogeneous Federated Learning [J].
Wang, Jianyu ;
Liu, Qinghua ;
Liang, Hao ;
Gauri, Joshi ;
Poor, H. Vincent .
IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2021, 69 :5234-5249
[45]   FedCos: A Scene-Adaptive Enhancement for Federated Learning [J].
Zhang, Hao ;
Wu, Tingting ;
Cheng, Siyao ;
Liu, Jie .
IEEE INTERNET OF THINGS JOURNAL, 2023, 10 (05) :4545-4556
[46]   FairGuard: A Fairness Attack and Defense Framework in Federated Learning [J].
Sheng, Xinyi ;
Yang, Zhengjie ;
Bao, Wei .
IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING, 2025, 22 (03) :2519-2532
[47]   Pruning-Based Adaptive Federated Learning at the Edge [J].
Yu, Dongxiao ;
Yuan, Yuan ;
Zou, Yifei ;
Zhang, Xiao ;
Liu, Yu ;
Cui, Lizhen ;
Cheng, Xiuzhen .
IEEE TRANSACTIONS ON COMPUTERS, 2025, 74 (05) :1538-1548
[48]   FedDPP: A Decentralized Peer-to-Peer Federated Learning Framework [J].
Gupta, Anupam ;
Mitra, Pabitra .
2024 5TH INTERNATIONAL CONFERENCE ON COMPUTERS AND ARTIFICIAL INTELLIGENCE TECHNOLOGY, CAIT, 2024, :481-486
[49]   HADFL: Heterogeneity-aware Decentralized Federated Learning Framework [J].
Cao, Jing ;
Lian, Zirui ;
Liu, Weihong ;
Zhu, Zongwei ;
Ji, Cheng .
2021 58TH ACM/IEEE DESIGN AUTOMATION CONFERENCE (DAC), 2021, :1-6
[50]   Decentralized federated learning based on blockchain: concepts, framework, and challenges [J].
Zhang, Haoran ;
Jiang, Shan ;
Xuan, Shichang .
COMPUTER COMMUNICATIONS, 2024, 216 :140-150