Heterogeneous Federated Learning: State-of-the-art and Research Challenges

Mang Ye, Xiuwen Fang, Bo Du, Pong C. Yuen, Dacheng Tao

Research output: Contribution to journalJournal articlepeer-review

2 Citations (Scopus)

Abstract

Federated learning (FL) has drawn increasing attention owing to its potential use in large-scale industrial applications. Existing FL works mainly focus on model homogeneous settings. However, practical FL typically faces the heterogeneity of data distributions, model architectures, network environments, and hardware devices among participant clients. Heterogeneous Federated Learning (HFL) is much more challenging, and corresponding solutions are diverse and complex. Therefore, a systematic survey on this topic about the research challenges and state-of-the-art is essential. In this survey, we firstly summarize the various research challenges in HFL from five aspects: statistical heterogeneity, model heterogeneity, communication heterogeneity, device heterogeneity, and additional challenges. In addition, recent advances in HFL are reviewed and a new taxonomy of existing HFL methods is proposed with an in-depth analysis of their pros and cons. We classify existing methods from three different levels according to the HFL procedure: data-level, model-level, and server-level. Finally, several critical and promising future research directions in HFL are discussed, which may facilitate further developments in this field. A periodically updated collection on HFL is available at https://github.com/marswhu/HFL_Survey.

Original languageEnglish
Article number79
Pages (from-to)1-44
Number of pages44
JournalACM Computing Surveys
Volume56
Issue number3
DOIs
Publication statusPublished - 21 Oct 2023

Scopus Subject Areas

  • Theoretical Computer Science
  • Computer Science(all)

User-Defined Keywords

  • Additional Key Words and PhrasesSurvey
  • federated learning
  • trustworthy AI

Fingerprint

Dive into the research topics of 'Heterogeneous Federated Learning: State-of-the-art and Research Challenges'. Together they form a unique fingerprint.

Cite this