Toward Memristive Phase‐Change Neural Network with High‐Quality Ultra‐Effective Highly‐Self‐Adjustable Online Learning

Abstract Memristive hardware with reconfigurable conductance levels are leading candidates for achieving artificial neural networks (ANNs). However, owing to difficulties in device character design and circuit combination, the ability to perform complicated online‐learning tasks on a memristive netw...

Full description

Bibliographic Details
Main Authors: Kian‐Guan Lim, Shao‐Xiang Go, Chun‐Chia Tan, Yu Jiang, Kui Cai, Tow‐Chong Chong, Stephen R. Elliott, Tae‐Hoon Lee, Desmond K. Loke
Format: Article
Language:English
Published: Wiley-VCH 2024-03-01
Series:Advanced Physics Research
Subjects:
Online Access:https://doi.org/10.1002/apxr.202300085
_version_ 1797270001078501376
author Kian‐Guan Lim
Shao‐Xiang Go
Chun‐Chia Tan
Yu Jiang
Kui Cai
Tow‐Chong Chong
Stephen R. Elliott
Tae‐Hoon Lee
Desmond K. Loke
author_facet Kian‐Guan Lim
Shao‐Xiang Go
Chun‐Chia Tan
Yu Jiang
Kui Cai
Tow‐Chong Chong
Stephen R. Elliott
Tae‐Hoon Lee
Desmond K. Loke
author_sort Kian‐Guan Lim
collection DOAJ
description Abstract Memristive hardware with reconfigurable conductance levels are leading candidates for achieving artificial neural networks (ANNs). However, owing to difficulties in device character design and circuit combination, the ability to perform complicated online‐learning tasks on a memristive network is not well understood. Here, tandem (T) material states are harnessed in a phase‐change memory (PCM) element, i.e., the primed‐amorphous state and the partial‐crystallized state, by utilizing an impetus‐and‐consequent pair pulse through a large degree of configurational ordering, and illustrate the development of an integrated system for achieving in‐memory computing and neural networks (NNs). A correct classification of 96.1% of 10,000 separate test images from the conventional Modified‐National‐Institute‐of‐Standards‐and‐Technology (MNIST) database in the tandem neural‐network (T‐NN) model is achieved, as well as image recognition for 28×28‐pixel pictures. The T‐NN configuration exhibits an in situ learning, with 50% of the elements stuck in the low‐conductance state, and at the same time, maintains an identification accuracy of ≈90%. The structural origin of the large degree of configurational‐ordering‐enhanced improvement in the extent of the conductance uniformity in the T‐based memristive element is revealed by theoretical studies. This work opens the door for attaining a widely relevant hardware system capable of performing artificial intelligence tasks with a large power‐time efficacy.
first_indexed 2024-04-25T01:57:19Z
format Article
id doaj.art-4cd0785974f34a4c991d68526e6d8b76
institution Directory Open Access Journal
issn 2751-1200
language English
last_indexed 2024-04-25T01:57:19Z
publishDate 2024-03-01
publisher Wiley-VCH
record_format Article
series Advanced Physics Research
spelling doaj.art-4cd0785974f34a4c991d68526e6d8b762024-03-07T15:43:41ZengWiley-VCHAdvanced Physics Research2751-12002024-03-0133n/an/a10.1002/apxr.202300085Toward Memristive Phase‐Change Neural Network with High‐Quality Ultra‐Effective Highly‐Self‐Adjustable Online LearningKian‐Guan Lim0Shao‐Xiang Go1Chun‐Chia Tan2Yu Jiang3Kui Cai4Tow‐Chong Chong5Stephen R. Elliott6Tae‐Hoon Lee7Desmond K. Loke8Department of Science, Mathematics and Technology Singapore University of Technology and Design Singapore 487372 SingaporeDepartment of Science, Mathematics and Technology Singapore University of Technology and Design Singapore 487372 SingaporeDepartment of Science, Mathematics and Technology Singapore University of Technology and Design Singapore 487372 SingaporeDepartment of Science, Mathematics and Technology Singapore University of Technology and Design Singapore 487372 SingaporeDepartment of Science, Mathematics and Technology Singapore University of Technology and Design Singapore 487372 SingaporeDepartment of Science, Mathematics and Technology Singapore University of Technology and Design Singapore 487372 SingaporePhysical and Theoretical Chemistry Laboratory University of Oxford Oxford OX1 3QZ UKSchool of Materials Science and Engineering Kyungpook National University Daegu 41566 Republic of KoreaDepartment of Science, Mathematics and Technology Singapore University of Technology and Design Singapore 487372 SingaporeAbstract Memristive hardware with reconfigurable conductance levels are leading candidates for achieving artificial neural networks (ANNs). However, owing to difficulties in device character design and circuit combination, the ability to perform complicated online‐learning tasks on a memristive network is not well understood. Here, tandem (T) material states are harnessed in a phase‐change memory (PCM) element, i.e., the primed‐amorphous state and the partial‐crystallized state, by utilizing an impetus‐and‐consequent pair pulse through a large degree of configurational ordering, and illustrate the development of an integrated system for achieving in‐memory computing and neural networks (NNs). A correct classification of 96.1% of 10,000 separate test images from the conventional Modified‐National‐Institute‐of‐Standards‐and‐Technology (MNIST) database in the tandem neural‐network (T‐NN) model is achieved, as well as image recognition for 28×28‐pixel pictures. The T‐NN configuration exhibits an in situ learning, with 50% of the elements stuck in the low‐conductance state, and at the same time, maintains an identification accuracy of ≈90%. The structural origin of the large degree of configurational‐ordering‐enhanced improvement in the extent of the conductance uniformity in the T‐based memristive element is revealed by theoretical studies. This work opens the door for attaining a widely relevant hardware system capable of performing artificial intelligence tasks with a large power‐time efficacy.https://doi.org/10.1002/apxr.202300085artificial neural networkimage classificationin situ online learningmemristive devicesphase‐change‐memory materials
spellingShingle Kian‐Guan Lim
Shao‐Xiang Go
Chun‐Chia Tan
Yu Jiang
Kui Cai
Tow‐Chong Chong
Stephen R. Elliott
Tae‐Hoon Lee
Desmond K. Loke
Toward Memristive Phase‐Change Neural Network with High‐Quality Ultra‐Effective Highly‐Self‐Adjustable Online Learning
Advanced Physics Research
artificial neural network
image classification
in situ online learning
memristive devices
phase‐change‐memory materials
title Toward Memristive Phase‐Change Neural Network with High‐Quality Ultra‐Effective Highly‐Self‐Adjustable Online Learning
title_full Toward Memristive Phase‐Change Neural Network with High‐Quality Ultra‐Effective Highly‐Self‐Adjustable Online Learning
title_fullStr Toward Memristive Phase‐Change Neural Network with High‐Quality Ultra‐Effective Highly‐Self‐Adjustable Online Learning
title_full_unstemmed Toward Memristive Phase‐Change Neural Network with High‐Quality Ultra‐Effective Highly‐Self‐Adjustable Online Learning
title_short Toward Memristive Phase‐Change Neural Network with High‐Quality Ultra‐Effective Highly‐Self‐Adjustable Online Learning
title_sort toward memristive phase change neural network with high quality ultra effective highly self adjustable online learning
topic artificial neural network
image classification
in situ online learning
memristive devices
phase‐change‐memory materials
url https://doi.org/10.1002/apxr.202300085
work_keys_str_mv AT kianguanlim towardmemristivephasechangeneuralnetworkwithhighqualityultraeffectivehighlyselfadjustableonlinelearning
AT shaoxianggo towardmemristivephasechangeneuralnetworkwithhighqualityultraeffectivehighlyselfadjustableonlinelearning
AT chunchiatan towardmemristivephasechangeneuralnetworkwithhighqualityultraeffectivehighlyselfadjustableonlinelearning
AT yujiang towardmemristivephasechangeneuralnetworkwithhighqualityultraeffectivehighlyselfadjustableonlinelearning
AT kuicai towardmemristivephasechangeneuralnetworkwithhighqualityultraeffectivehighlyselfadjustableonlinelearning
AT towchongchong towardmemristivephasechangeneuralnetworkwithhighqualityultraeffectivehighlyselfadjustableonlinelearning
AT stephenrelliott towardmemristivephasechangeneuralnetworkwithhighqualityultraeffectivehighlyselfadjustableonlinelearning
AT taehoonlee towardmemristivephasechangeneuralnetworkwithhighqualityultraeffectivehighlyselfadjustableonlinelearning
AT desmondkloke towardmemristivephasechangeneuralnetworkwithhighqualityultraeffectivehighlyselfadjustableonlinelearning