Integrated Artificial Neural Network with Trainable Activation Function Enabled by Topological Insulator-Based Spin-Orbit Torque Devices

doi:10.1021/acsnano.4c03278

	Integrated Artificial Neural Network with Trainable Activation Function Enabled by Topological Insulator-Based Spin-Orbit Torque Devices
	Huang, Puyang1 ; Liu, Xinqi2,3 ; Xin, Yue1 ; Gu, Yu1 ; Lee, Albert 4; Zhang, Yifan1 ; Xu, Zhuo1 ; Chen, Peng1 ; Zhang, Yu 5; Deng, Weijie2 ; Yu, Guoqiang 5; Wu, Di 4; Liu, Zhongkai2,3 ; Yao, Qi2,3 ; Yang, Yumeng1 ; Zhu, Zhifeng1 ; Kou, Xufeng1,3
	2024-10-01
发表期刊	ACS NANO (IF:15.8[JCR-2023],16.2[5-Year])
ISSN	1936-0851
EISSN	1936-086X
卷号	18 期号:43
发表状态	已发表
DOI	10.1021/acsnano.4c03278
摘要	Nonvolatile memristors offer a salient platform for artificial neural network (ANN), yet the integration of different function and algorithm blocks into one hardware system remains challenging. Here we demonstrate the brain-like synaptic (SOT-S) and neuronal (SOT-N) functions in the Bi2Te3/CrTe2 heterostructure-based spin-orbit torque (SOT) device. The SOT-S unit exhibits highly linear and symmetrical long-term potentiation/depression process, resulting in a fast-training of the MNIST data set with the classification accuracy above 90%. Meanwhile, the Sigmoid-shape transition curve inherited in the SOT-N cell replaces the software-based activation function block, hence reducing the system complexity. On this basis, we employ a serial-connected, voltage-mode sensing ANN architecture to enhance the vector-matrix multiplication signal strength with low reading error of 0.61% while simplifying the peripheral circuitry. Furthermore, the trainable activation function of SOT-N enables the implementation of the Batch Normalization algorithm and activation operation within one clock cycle, which bring about improved on/off-chip training performance close to the ideal baseline.
关键词	spin-orbit torque anomalous Hall effect topological insulator van der Waals ferromagnet artificial neural network
URL	查看原文
收录类别	SCI ; EI
语种	英语
资助项目	National Key R&D Program of China[2021YFA0715503] ; National Natural Science Foundation of China[92164104] ; Major Project of Shanghai Municipal Science and Technology[2018SHZDZX02] ; ShanghaiTech Material Device and Soft Matter Nanofabrication Laboratories[SMN180827] ; Shanghai Rising-Star Program[21QA1406000]
WOS研究方向	Chemistry ; Science & Technology - Other Topics ; Materials Science
WOS类目	Chemistry, Multidisciplinary ; Chemistry, Physical ; Nanoscience & Nanotechnology ; Materials Science, Multidisciplinary
WOS记录号	WOS:001337593800001
出版者	AMER CHEMICAL SOC
EI入藏号	20244317232684
EI主题词	Van der Waals forces
EI分类号	1004 ; 103 ; 1106.3.1 ; 1301.1.3 ; 405.3 Surveying ; 701.2 Magnetism: Basic Concepts and Phenomena ; 708.1 Dielectric Materials ; 713.5 Electronic Circuits Other Than Amplifiers, Oscillators, Modulators, Limiters, Discriminators or Mixers ; 715 Electronic Equipment, General Purpose and Industrial ; 742.1 Photography ; 801.3 Colloid Chemistry
原始文献类型	Article in Press
引用统计	正在获取...
文献类型	期刊论文
条目标识符	https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/329087
专题	物质科学与技术学院信息科学与技术学院物质科学与技术学院_PI研究组_柳仲楷组信息科学与技术学院_PI研究组_寇煦丰组物质科学与技术学院_博士生信息科学与技术学院_硕士生信息科学与技术学院_博士生物质科学与技术学院_公共科研平台_拓扑物理实验室信息科学与技术学院_PI研究组_祝智峰组信息科学与技术学院_PI研究组_杨雨梦组
通讯作者	Zhu, Zhifeng; Kou, Xufeng
作者单位	1.Shanghai Tech Univ, Sch Informat Sci & Technol, Shanghai 201210, Peoples R China 2.Shanghai Tech Univ, Sch Phys Sci & Technol, Shanghai 201210, Peoples R China 3.Shanghai Tech Univ, Shanghai Tech Lab Topol Phys, Shanghai 201210, Peoples R China 4.Suzhou Inston Technol Co Ltd, Suzhou 215121, Jiangsu, Peoples R China 5.Chinese Acad Sci, Phys Inst Phys, Beijing Natl Lab Condensed Matter, Beijing 100190, Peoples R China
第一作者单位	信息科学与技术学院
通讯作者单位	信息科学与技术学院; 上海科技大学
第一作者的第一单位	信息科学与技术学院
推荐引用方式 GB/T 7714	Huang, Puyang,Liu, Xinqi,Xin, Yue,et al. Integrated Artificial Neural Network with Trainable Activation Function Enabled by Topological Insulator-Based Spin-Orbit Torque Devices[J]. ACS NANO,2024,18(43).
APA	Huang, Puyang.,Liu, Xinqi.,Xin, Yue.,Gu, Yu.,Lee, Albert.,...&Kou, Xufeng.(2024).Integrated Artificial Neural Network with Trainable Activation Function Enabled by Topological Insulator-Based Spin-Orbit Torque Devices.ACS NANO,18(43).
MLA	Huang, Puyang,et al."Integrated Artificial Neural Network with Trainable Activation Function Enabled by Topological Insulator-Based Spin-Orbit Torque Devices".ACS NANO 18.43(2024).