Semi-asynchronous federated learning algorithm based on clustering and training efficiency

Jiayi Lan; Zhenrong Zhang

doi:10.1117/12.2641418

4 August 2022 Semi-asynchronous federated learning algorithm based on clustering and training efficiency

Jiayi Lan, Zhenrong Zhang

Proceedings Volume 12306, Second International Conference on Digital Signal and Computer Communications (DSCC 2022); 123060X (2022) https://doi.org/10.1117/12.2641418
Event: Second International Conference on Digital Signal and Computer Communications (DSCC 2022), 2022, Changchun, China

Abstract

In order to solve the problem of device heterogeneity in federated learning, FedAsync proposes asynchronous federated learning, that is, the server and the client interact in an asynchronous manner, that is, the server updates the global model immediately after receiving the local model. Communication between server and client is non-blocking. Therefore, the server and client can update the model at any time without synchronization, which is advantageous when the devices have heterogeneous conditions. However, when the gap between clients increases, the time difference of the uploaded model parameters becomes larger, which has a great impact on the accuracy. In response to the above problems, this paper first introduces the concept of training efficiency, evaluates the training ability of customers, and determines the training frequency of customers according to the training ability, thereby narrowing the gap between customers. Second, change the semi-asynchronous strategy, that is, the uploaded model is not updated at any time, but aggregated and updated after a certain amount or a certain time is satisfied. Finally, the effect is verified by simulation experiments.

Citation Download Citation

Jiayi Lan and Zhenrong Zhang "Semi-asynchronous federated learning algorithm based on clustering and training efficiency", Proc. SPIE 12306, Second International Conference on Digital Signal and Computer Communications (DSCC 2022), 123060X (4 August 2022); https://doi.org/10.1117/12.2641418

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

Performance modeling

Data modeling

Instrument modeling

Computer engineering

Computer science

Computer simulations

Software development

Keywords/Phrases

Search In:

Publication Years