Hierarchical federated learning algorithm based on efficient aggregation mechanism

Fangjing Li; Wenbo Zhang; Hongbo Zhu

doi:10.1117/12.2675127

25 May 2023 Hierarchical federated learning algorithm based on efficient aggregation mechanism

Fangjing Li, Wenbo Zhang, Hongbo Zhu

Proceedings Volume 12636, Third International Conference on Machine Learning and Computer Application (ICMLCA 2022); 1263654 (2023) https://doi.org/10.1117/12.2675127
Event: Third International Conference on Machine Learning and Computer Application (ICMLCA 2022), 2022, Shenyang, China

Abstract

The existing federated learning algorithm is affected by the heterogeneity of multiple clients and the frequent information interactions between clients and servers, resulting in low training efficiency. Therefore, the edge servers are introduced to the client and the center server in this paper, in order to construct a hierarchical federated learning framework that can effectively reduce the frequency and cost of communication between the client and the center server, which also improves the communication efficiency. At the same time, an efficient time-based hierarchical aggregation mechanism is designed for the hierarchical federated learning framework. The edge server receives local models within a specified period of time and achieves dynamic weighted aggregation according to the obsolescence of the local models obtained in each round, so as to improve the overall training efficiency and ensure the accuracy of the models. In experiments, it is shown that compared with the existing federated learning framework, the hierarchical federated learning framework based on efficient aggregation mechanism proposed in this paper can significantly improve the training efficiency on both MNIST and CIFAR-10 datasets.

Citation Download Citation

Fangjing Li, Wenbo Zhang, and Hongbo Zhu "Hierarchical federated learning algorithm based on efficient aggregation mechanism", Proc. SPIE 12636, Third International Conference on Machine Learning and Computer Application (ICMLCA 2022), 1263654 (25 May 2023); https://doi.org/10.1117/12.2675127

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
6 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Education and training

Machine learning

Data modeling

Instrument modeling

Computer simulations

Data privacy

Data communications

Show All Keywords

Keywords/Phrases

Search In:

Publication Years