A Federated Learning Client Selection Method via Multi-Task Deep Reinforcement Learning

Le Hou, Laisen Nie, Xinyang Deng

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Federated Learning (FL) is a privacy-preserving paradigm for training machine learning (ML) models, crucial for data privacy and security protection.It has garnered significant attention from both industry and academia.Typically, clients are selected randomly for training and model aggregation in FL scenarios.However, heterogeneity in data distribution and hardware among devices leads to problems such as slow model convergence, low accuracy, and high computational overhead.To address the issues of statistical heterogeneity and system heterogeneity in FL, this paper proposes an intelligent client selection framework via multi-task deep reinforcement learning (DRL).Additionally, two reward functions are introduced to alleviate the heterogeneity problem by maximizing model performance and minimizing system latency.Experimental results on MNIST and CIFAR-10 datasets demonstrate the effectiveness of the proposed method.

Original languageEnglish
Title of host publicationProceedings of 2024 IEEE International Conference on Unmanned Systems, ICUS 2024
EditorsRong Song
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1357-1362
Number of pages6
ISBN (Electronic)9798350384185
DOIs
StatePublished - 2024
Event2024 IEEE International Conference on Unmanned Systems, ICUS 2024 - Nanjing, China
Duration: 18 Oct 202420 Oct 2024

Publication series

NameProceedings of 2024 IEEE International Conference on Unmanned Systems, ICUS 2024

Conference

Conference2024 IEEE International Conference on Unmanned Systems, ICUS 2024
Country/TerritoryChina
CityNanjing
Period18/10/2420/10/24

Keywords

  • client selection
  • federated learning
  • multi-task deep reinforcement learning
  • statistical heterogeneity
  • system heterogeneity

Fingerprint

Dive into the research topics of 'A Federated Learning Client Selection Method via Multi-Task Deep Reinforcement Learning'. Together they form a unique fingerprint.

Cite this