Skip to main navigation Skip to search Skip to main content

Asynchronous Federated Learning Framework Based on Dynamic Selective Transmission

  • Ruizhuo Zhang
  • , Wenjian Luo*
  • , Yongkang Luo
  • , Shaocong Xue
  • *Corresponding author for this work
  • School of Computer Science and Technology, Harbin Institute of Technology
  • Peng Cheng Laboratory

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

Abstract

This paper proposes an asynchronous federated learning framework based on dynamic selective transmission to solve the communication efficiency problem in asynchronous federated learning. The framework first dynamically determines the network layer to be transmitted in each round according to the training progress, and then dynamically adjusts the ratio of parameters to be transmitted according to the training progress and model staleness for the selected network layers, which effectively reduces the negative impact of the local model staleness of the client on the performance of global model while significantly reducing the uplink communication cost of the asynchronous federated learning framework. On two datasets, this paper designs experiments to compare the proposed framework with the existing work. Through the analysis of the experimental results, the effectiveness of the proposed framework is verified.

Original languageEnglish
Title of host publicationAdvances in Swarm Intelligence - 14th International Conference, ICSI 2023, Proceedings
EditorsYing Tan, Yuhui Shi, Wenjian Luo
PublisherSpringer Science and Business Media Deutschland GmbH
Pages193-203
Number of pages11
ISBN (Print)9783031366246
DOIs
StatePublished - 2023
Externally publishedYes
Event14th International Conference on Advances in Swarm Intelligence, ICSI 2023 - Shenzhen, China
Duration: 14 Jul 202318 Jul 2023

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13969 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference14th International Conference on Advances in Swarm Intelligence, ICSI 2023
Country/TerritoryChina
CityShenzhen
Period14/07/2318/07/23

Keywords

  • Asynchronous federated learning
  • Communication efficiency
  • Federated learning
  • Machine learning

Fingerprint

Dive into the research topics of 'Asynchronous Federated Learning Framework Based on Dynamic Selective Transmission'. Together they form a unique fingerprint.

Cite this