Publication
Title
SMART : selective MAC zero-optimization for neural network reliability under radiation
Author
Abstract
Neural networks running on low-power edge devices can help in achieving ubiquitous computing with limited infrastructure. When such edge devices are deployed in conventional and extreme environments without the necessary shielding, they must be fault tolerant for reliable operation. As a pilot study, we focused on embedding fault tolerance into neural networks by proposing a novel selective multiply-accumulate zerooptimization technique based on whether the value of an input provided to a neuron of a neural network is zero. If the value is zero, then the corresponding multiply-accumulate operation is bypassed. We subjected the operating system-based implementation of the optimization technique to radiation test campaigns using approximately 14 MeV neutrons, and found the proposed optimization technique to improve the fault tolerance of the tested neural network by approximately 44%.
Language
English
Source (journal)
Microelectronics and reliability. - Oxford
Publication
Oxford : 2023
ISSN
0026-2714
DOI
10.1016/J.MICROREL.2023.115092
Volume/pages
150 (2023) , 7 p.
Article Reference
115092
ISI
001106952400001
Full text (Publisher's DOI)
Full text (open access)
UAntwerpen
Faculty/Department
Research group
Publication type
Subject
Affiliation
Publications with a UAntwerp address
External links
Web of Science
Record
Identifier
Creation 09.01.2024
Last edited 11.01.2024
To cite this reference