FL-SGCN: federated learning on spiking graph convolutional networks

Yuan Wang; Yi Liu; Hongyu Zhao

doi:10.1117/12.2671341

29 March 2023 FL-SGCN: federated learning on spiking graph convolutional networks

Yuan Wang, Yi Liu, Hongyu Zhao

Proceedings Volume 12594, Second International Conference on Electronic Information Engineering and Computer Communication (EIECC 2022); 125942Q (2023) https://doi.org/10.1117/12.2671341
Event: Second International Conference on Electronic Information Engineering and Computer Communication (EIECC 2022), 2022, Xi'an, China

Abstract

Graph Convolutional Networks (GCNs) are neural models that capture the dependence of graphs via message passing between the nodes of graphs. However, SGD-based GCNs training algorithms require a lot of computing resources, which limits the application scenarios of GCNs. Spiking neural networks (SNNs) are an efficient and energy-saving network with strong biological interpretation and great development prospects, helping GCNs work on mobile devices with strict power limits. However, the training effect of SNNs is still low compare to traditional CNN networks. In this paper, we combine federated learning (FL) structure with spiking GCN networks, and introduce asynchronous architecture to improve the training effect and improve the robustness of the system. The digital simulation experiment proves our conclusion.

Citation Download Citation

Yuan Wang, Yi Liu, and Hongyu Zhao "FL-SGCN: federated learning on spiking graph convolutional networks", Proc. SPIE 12594, Second International Conference on Electronic Information Engineering and Computer Communication (EIECC 2022), 125942Q (29 March 2023); https://doi.org/10.1117/12.2671341

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