Evolutionary Multi-model Federated Learning on Malicious and Heterogeneous Data

被引:2
|
作者
Shang, Chikai [1 ]
Gu, Fangqing [1 ]
Jiang, Jiaqi [1 ]
机构
[1] Guangdong Univ Technol, Sch Math & Stat, Guangzhou, Guangdong, Peoples R China
来源
2023 23RD IEEE INTERNATIONAL CONFERENCE ON DATA MINING WORKSHOPS, ICDMW 2023 | 2023年
关键词
Federated learning; particle swarm optimization; heterogeneous data; malicious data;
D O I
10.1109/ICDMW60847.2023.00055
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
The federated learning system is vulnerable to data poisoning attacks from malicious clients, who often present heterogeneous data. Poisoning such heterogeneous data can lead to significant diversity in the local datasets. Traditional singlemodel federated learning struggles to effectively handle this diverse knowledge. To address this challenge, we propose an evolutionary multi-model federated learning framework called EMFL, which trains multiple models with diversity to learn from the diverse data effectively. Specifically, we introduce a gradientbased particle swarm optimization approach, which incorporates gradient information from the models and utilizes a multielite strategy to achieve quick convergence and promote model diversity. Furthermore, we propose a client-matching strategy based on cosine similarity of multi-model updates, allowing for analyzing the similarity of distribution patterns in diverse data and matching clients with concentrated distributions, thus helping to prevent excessive models from converging to malicious data and causing extreme model diversity. The proposed EMFL employs multiple lightweight models instead of a large architecture one to overcome training challenges for clients with limited computing power. Experimental results on nine datasets demonstrate consistent outperformance of our proposed algorithm compared to its counterparts.
引用
收藏
页码:386 / 395
页数:10
相关论文
共 50 条
  • [41] Towards Accelerating the Adoption of Federated Learning for Heterogeneous Data
    Ntokos, Christos
    Bakalos, Nikolaos
    Kalogeras, Dimitrios
    PROCEEDINGS OF THE 16TH ACM INTERNATIONAL CONFERENCE ON PERVASIVE TECHNOLOGIES RELATED TO ASSISTIVE ENVIRONMENTS, PETRA 2023, 2023, : 617 - 624
  • [42] Efficient Hierarchical Personalized Federated Learning on Heterogeneous Data
    Wang, Sai
    You, Changsheng
    Zhao, Junhui
    Gong, Yi
    2024 IEEE/CIC INTERNATIONAL CONFERENCE ON COMMUNICATIONS IN CHINA, ICCC, 2024,
  • [43] Data-Heterogeneous Hierarchical Federated Learning with Mobility
    Chen, Tan
    Yan, Jintao
    Sun, Yuxuan
    Zhou, Sheng
    Gunduz, Deniz
    Niu, Zhisheng
    2023 35TH INTERNATIONAL TELETRAFFIC CONGRESS, ITC-35, 2023,
  • [44] Categorical Management of Multi-Model Data
    Svoboda, Martin
    Contos, Pavel
    Holubova, Irena
    IDEAS 2021: 25TH INTERNATIONAL DATABASE ENGINEERING & APPLICATIONS SYMPOSIUM, 2021, : 134 - 140
  • [45] Robust Heterogeneous Federated Learning under Data Corruption
    Fang, Xiuwen
    Ye, Mang
    Yang, Xiyuan
    2023 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION, ICCV, 2023, : 4997 - 5007
  • [46] Convergence Analysis of Sequential Federated Learning on Heterogeneous Data
    Li, Yipeng
    Lyu, Xinchen
    ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 36 (NEURIPS 2023), 2023,
  • [47] Multi-model classification method in heterogeneous image databases
    Kachouri, Rostom
    Djemal, Khalifa
    Maaref, Hichem
    PATTERN RECOGNITION, 2010, 43 (12) : 4077 - 4088
  • [48] Unified Views for Querying Heterogeneous Multi-model Polystores
    El Ahdab, Lea
    Teste, Olivier
    Megdiche, Imen
    Peninou, Andre
    BIG DATA ANALYTICS AND KNOWLEDGE DISCOVERY, DAWAK 2023, 2023, 14148 : 319 - 324
  • [49] Towards Data-Independent Knowledge Transfer in Model-Heterogeneous Federated Learning
    Zhang, Jie
    Guo, Song
    Guo, Jingcai
    Zeng, Deze
    Zhou, Jingren
    Zomaya, Albert Y.
    IEEE TRANSACTIONS ON COMPUTERS, 2023, 72 (10) : 2888 - 2901
  • [50] FedCav: Contribution-aware Model Aggregation on Distributed Heterogeneous Data in Federated Learning
    Zeng, Hui
    Zhou, Tongqing
    Guo, Yeting
    Cai, Zhiping
    Liu, Fang
    50TH INTERNATIONAL CONFERENCE ON PARALLEL PROCESSING, 2021,