Summary

International Symposium on Nonlinear Theory and Its Applications

2022

Session Number:A4L-D

Session:

Number:A4L-D-03

Prospects of Energy-Efficient Edge-AI Accelerator Architecture Using Nonvolatile Logic

Masanori Natsui ,   Daisuke Suzuki ,   Yasuhiro Takako ,   Akira Tamakoshi ,   Takahiro Hanyu,  

pp.136-138

Publication Date:12/12/2022

Online ISSN:2188-5079

DOI:10.34385/proc.71.A4L-D-03

PDF download (534KB)

Summary:
The realization of energy-efficient edge-AI hardware is an important issue for utilizing it as a fundamental technology for the next-generation IoT society. Nonvolatile logic-circuit technology based on MTJ devices is a key to solving this challenge. In this paper, we discuss design guidelines and prospects for edge-AI hardware that enables energy-saving operation by maximizing the use of nonvolatile memory functions.