Graphene-based 3D XNOR-VRRAM with ternary precision for neuromorphic computing

Abstract Recent studies on neural network quantization have demonstrated a beneficial compromise between accuracy, computation rate, and architecture size. Implementing a 3D Vertical RRAM (VRRAM) array accompanied by device scaling may further improve such networks’ density and energy consumption. I...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Batyrbek Alimkhanuly, Joon Sohn, Ik-Joon Chang, Seunghyun Lee
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2021
Materias:	Materials of engineering and construction. Mechanics of materials TA401-492 Chemistry QD1-999
Acceso en línea:	https://doaj.org/article/9403e82d0eff4fdfb77920892c449198
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:9403e82d0eff4fdfb77920892c449198
record_format	dspace
spelling	oai:doaj.org-article:9403e82d0eff4fdfb77920892c4491982021-12-02T17:15:58ZGraphene-based 3D XNOR-VRRAM with ternary precision for neuromorphic computing10.1038/s41699-021-00236-x2397-7132https://doaj.org/article/9403e82d0eff4fdfb77920892c4491982021-05-01T00:00:00Zhttps://doi.org/10.1038/s41699-021-00236-xhttps://doaj.org/toc/2397-7132Abstract Recent studies on neural network quantization have demonstrated a beneficial compromise between accuracy, computation rate, and architecture size. Implementing a 3D Vertical RRAM (VRRAM) array accompanied by device scaling may further improve such networks’ density and energy consumption. Individual device design, optimized interconnects, and careful material selection are key factors determining the overall computation performance. In this work, the impact of replacing conventional devices with microfabricated, graphene-based VRRAM is investigated for circuit and algorithmic levels. By exploiting a sub-nm thin 2D material, the VRRAM array demonstrates an improved read/write margins and read inaccuracy level for the weighted-sum procedure. Moreover, energy consumption is significantly reduced in array programming operations. Finally, an XNOR logic-inspired architecture designed to integrate 1-bit ternary precision synaptic weights into graphene-based VRRAM is introduced. Simulations on VRRAM with metal and graphene word-planes demonstrate 83.5 and 94.1% recognition accuracy, respectively, denoting the importance of material innovation in neuromorphic computing.Batyrbek AlimkhanulyJoon SohnIk-Joon ChangSeunghyun LeeNature PortfolioarticleMaterials of engineering and construction. Mechanics of materialsTA401-492ChemistryQD1-999ENnpj 2D Materials and Applications, Vol 5, Iss 1, Pp 1-10 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Materials of engineering and construction. Mechanics of materials TA401-492 Chemistry QD1-999
spellingShingle	Materials of engineering and construction. Mechanics of materials TA401-492 Chemistry QD1-999 Batyrbek Alimkhanuly Joon Sohn Ik-Joon Chang Seunghyun Lee Graphene-based 3D XNOR-VRRAM with ternary precision for neuromorphic computing
description	Abstract Recent studies on neural network quantization have demonstrated a beneficial compromise between accuracy, computation rate, and architecture size. Implementing a 3D Vertical RRAM (VRRAM) array accompanied by device scaling may further improve such networks’ density and energy consumption. Individual device design, optimized interconnects, and careful material selection are key factors determining the overall computation performance. In this work, the impact of replacing conventional devices with microfabricated, graphene-based VRRAM is investigated for circuit and algorithmic levels. By exploiting a sub-nm thin 2D material, the VRRAM array demonstrates an improved read/write margins and read inaccuracy level for the weighted-sum procedure. Moreover, energy consumption is significantly reduced in array programming operations. Finally, an XNOR logic-inspired architecture designed to integrate 1-bit ternary precision synaptic weights into graphene-based VRRAM is introduced. Simulations on VRRAM with metal and graphene word-planes demonstrate 83.5 and 94.1% recognition accuracy, respectively, denoting the importance of material innovation in neuromorphic computing.
format	article
author	Batyrbek Alimkhanuly Joon Sohn Ik-Joon Chang Seunghyun Lee
author_facet	Batyrbek Alimkhanuly Joon Sohn Ik-Joon Chang Seunghyun Lee
author_sort	Batyrbek Alimkhanuly
title	Graphene-based 3D XNOR-VRRAM with ternary precision for neuromorphic computing
title_short	Graphene-based 3D XNOR-VRRAM with ternary precision for neuromorphic computing
title_full	Graphene-based 3D XNOR-VRRAM with ternary precision for neuromorphic computing
title_fullStr	Graphene-based 3D XNOR-VRRAM with ternary precision for neuromorphic computing
title_full_unstemmed	Graphene-based 3D XNOR-VRRAM with ternary precision for neuromorphic computing
title_sort	graphene-based 3d xnor-vrram with ternary precision for neuromorphic computing
publisher	Nature Portfolio
publishDate	2021
url	https://doaj.org/article/9403e82d0eff4fdfb77920892c449198
work_keys_str_mv	AT batyrbekalimkhanuly graphenebased3dxnorvrramwithternaryprecisionforneuromorphiccomputing AT joonsohn graphenebased3dxnorvrramwithternaryprecisionforneuromorphiccomputing AT ikjoonchang graphenebased3dxnorvrramwithternaryprecisionforneuromorphiccomputing AT seunghyunlee graphenebased3dxnorvrramwithternaryprecisionforneuromorphiccomputing
_version_	1718381214692278272

Graphene-based 3D XNOR-VRRAM with ternary precision for neuromorphic computing

Ejemplares similares