Optimising the computational and cost efficiency of hierarchical federated edge learning

Chen Wang

doi:10.1117/12.3025935

6 May 2024 Optimising the computational and cost efficiency of hierarchical federated edge learning

Chen Wang

Proceedings Volume 13161, Fourth International Conference on Telecommunications, Optics, and Computer Science (TOCS 2023); 131610Q (2024) https://doi.org/10.1117/12.3025935
Event: Fourth International Conference on Telecommunications, Optics and Computer Science (TOCS 2023), 2023, Xi’an, China

Abstract

Hierarchical Federated Edge Learning (HFEL) is a promising and efficient framework , providing privacy preservation, which aims to address the issue of limited resources and network congestion by utilizing the available resources in the edge network. This paper proposes a new scheme, HFEL-Q, based on HFEL to address the high energy consumption problem of FL training, as well as the inherent communication and user heterogeneity problems. To improve training performance, a utility function is designed based on users' learning quality and training time to efficiently select the user group with the highest utility. Additionally, a frequency determination method is employed to optimize idle time and reduce energy consumption during training. Finally, the performance of HFEL-Q is evaluated on two real datasets to demonstrate its superiority over state-of-the-art baselines in terms of training rate, accuracy, and energy savings.

(2024) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Chen Wang "Optimising the computational and cost efficiency of hierarchical federated edge learning", Proc. SPIE 13161, Fourth International Conference on Telecommunications, Optics, and Computer Science (TOCS 2023), 131610Q (6 May 2024); https://doi.org/10.1117/12.3025935

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