基于STDP的脉冲神经网络片上无监督学习

IF 2.1 4区 工程技术 Q3 ENGINEERING, ELECTRICAL & ELECTRONIC IEEE Transactions on Nanotechnology Pub Date : 2023-07-06 DOI:10.1109/TNANO.2023.3293011
Abhinav Gupta;Sneh Saurabh
{"title":"基于STDP的脉冲神经网络片上无监督学习","authors":"Abhinav Gupta;Sneh Saurabh","doi":"10.1109/TNANO.2023.3293011","DOIUrl":null,"url":null,"abstract":"In this article, we propose an energy-efficient Ge-based device that enables on-chip unsupervised learning using Spike-Timing-Dependent-Plasticity (STDP) in a Spiking Neural Network (SNN). A Ferromagnetic Domain Wall (FM-DW) based device, which has decoupled read and write paths, is used as a synapse in this work. The proposed device comprises a dual pocket Fully-Depleted Silicon-on-Insulator (FD-SOI) MOSFET with dual asymmetric gates. Using a well-calibrated 2D device simulation model, we show that a pair of such devices can generate a current, which depends exponentially on the temporal correlation of spiking events in the pre- and post-neuronal layer. This current is fed to the FM-DW synapse, which in turn modulates the conductance of the synapse in accordance with the STDP learning rule. The proposed implementation requires 2-3× fewer transistors and offers a lower latency compared to existing literature. We further demonstrate the application of the proposed device at the system-level to train an SNN to recognize handwritten digits in the MNIST dataset and obtained a classification accuracy of 84%.","PeriodicalId":449,"journal":{"name":"IEEE Transactions on Nanotechnology","volume":"22 ","pages":"365-376"},"PeriodicalIF":2.1000,"publicationDate":"2023-07-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"On-Chip Unsupervised Learning Using STDP in a Spiking Neural Network\",\"authors\":\"Abhinav Gupta;Sneh Saurabh\",\"doi\":\"10.1109/TNANO.2023.3293011\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"In this article, we propose an energy-efficient Ge-based device that enables on-chip unsupervised learning using Spike-Timing-Dependent-Plasticity (STDP) in a Spiking Neural Network (SNN). A Ferromagnetic Domain Wall (FM-DW) based device, which has decoupled read and write paths, is used as a synapse in this work. The proposed device comprises a dual pocket Fully-Depleted Silicon-on-Insulator (FD-SOI) MOSFET with dual asymmetric gates. Using a well-calibrated 2D device simulation model, we show that a pair of such devices can generate a current, which depends exponentially on the temporal correlation of spiking events in the pre- and post-neuronal layer. This current is fed to the FM-DW synapse, which in turn modulates the conductance of the synapse in accordance with the STDP learning rule. The proposed implementation requires 2-3× fewer transistors and offers a lower latency compared to existing literature. We further demonstrate the application of the proposed device at the system-level to train an SNN to recognize handwritten digits in the MNIST dataset and obtained a classification accuracy of 84%.\",\"PeriodicalId\":449,\"journal\":{\"name\":\"IEEE Transactions on Nanotechnology\",\"volume\":\"22 \",\"pages\":\"365-376\"},\"PeriodicalIF\":2.1000,\"publicationDate\":\"2023-07-06\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"IEEE Transactions on Nanotechnology\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://ieeexplore.ieee.org/document/10174725/\",\"RegionNum\":4,\"RegionCategory\":\"工程技术\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"ENGINEERING, ELECTRICAL & ELECTRONIC\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE Transactions on Nanotechnology","FirstCategoryId":"5","ListUrlMain":"https://ieeexplore.ieee.org/document/10174725/","RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}
引用次数: 0

摘要

在本文中,我们提出了一种节能的基于ge的设备,该设备在尖峰神经网络(SNN)中使用尖峰时间依赖可塑性(STDP)实现片上无监督学习。在这项工作中,使用了一种基于铁磁畴壁(FM-DW)的器件作为突触,该器件具有解耦的读写路径。所提出的器件包括具有双非对称栅极的双口袋完全耗尽绝缘体上硅(FD-SOI) MOSFET。使用校准良好的二维器件模拟模型,我们表明一对这样的器件可以产生电流,其指数依赖于神经元层前和神经元层后尖峰事件的时间相关性。该电流被馈送到FM-DW突触,该突触反过来根据STDP学习规则调节突触的电导。与现有文献相比,提出的实现需要的晶体管减少了2-3倍,并且提供了更低的延迟。我们进一步演示了该装置在系统级的应用,以训练SNN识别MNIST数据集中的手写数字,并获得了84%的分类准确率。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
On-Chip Unsupervised Learning Using STDP in a Spiking Neural Network
In this article, we propose an energy-efficient Ge-based device that enables on-chip unsupervised learning using Spike-Timing-Dependent-Plasticity (STDP) in a Spiking Neural Network (SNN). A Ferromagnetic Domain Wall (FM-DW) based device, which has decoupled read and write paths, is used as a synapse in this work. The proposed device comprises a dual pocket Fully-Depleted Silicon-on-Insulator (FD-SOI) MOSFET with dual asymmetric gates. Using a well-calibrated 2D device simulation model, we show that a pair of such devices can generate a current, which depends exponentially on the temporal correlation of spiking events in the pre- and post-neuronal layer. This current is fed to the FM-DW synapse, which in turn modulates the conductance of the synapse in accordance with the STDP learning rule. The proposed implementation requires 2-3× fewer transistors and offers a lower latency compared to existing literature. We further demonstrate the application of the proposed device at the system-level to train an SNN to recognize handwritten digits in the MNIST dataset and obtained a classification accuracy of 84%.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
IEEE Transactions on Nanotechnology
IEEE Transactions on Nanotechnology 工程技术-材料科学:综合
CiteScore
4.80
自引率
8.30%
发文量
74
审稿时长
8.3 months
期刊介绍: The IEEE Transactions on Nanotechnology is devoted to the publication of manuscripts of archival value in the general area of nanotechnology, which is rapidly emerging as one of the fastest growing and most promising new technological developments for the next generation and beyond.
期刊最新文献
High-Speed and Area-Efficient Serial IMPLY-Based Approximate Subtractor and Comparator for Image Processing and Neural Networks Design of a Graphene Based Terahertz Perfect Metamaterial Absorber With Multiple Sensing Performance Impact of Electron and Hole Trap Profiles in BE-TOX on Retention Characteristics of 3D NAND Flash Memory Full 3-D Monte Carlo Simulation of Coupled Electron-Phonon Transport: Self-Heating in a Nanoscale FinFET Experimental Investigations and Characterization of Surfactant Activated Mixed Metal Oxide (MMO) Nanomaterial
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1