Robust high-dimensional memory-augmented neural networks

The implementation of memory-augmented neural networks using conventional computer architectures is challenging due to a large number of read and write operations. Here, Karunaratne, Schmuck et al. propose an architecture that enables analog in-memory computing on high-dimensional vectors at accurac...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Geethan Karunaratne, Manuel Schmuck, Manuel Le Gallo, Giovanni Cherubini, Luca Benini, Abu Sebastian, Abbas Rahimi
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2021
Materias:	Science Q
Acceso en línea:	https://doaj.org/article/b553751b28d34003af781f5a72f0719a
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:b553751b28d34003af781f5a72f0719a
record_format	dspace
spelling	oai:doaj.org-article:b553751b28d34003af781f5a72f0719a2021-12-02T16:55:32ZRobust high-dimensional memory-augmented neural networks10.1038/s41467-021-22364-02041-1723https://doaj.org/article/b553751b28d34003af781f5a72f0719a2021-04-01T00:00:00Zhttps://doi.org/10.1038/s41467-021-22364-0https://doaj.org/toc/2041-1723The implementation of memory-augmented neural networks using conventional computer architectures is challenging due to a large number of read and write operations. Here, Karunaratne, Schmuck et al. propose an architecture that enables analog in-memory computing on high-dimensional vectors at accuracy matching 32-bit software equivalent.Geethan KarunaratneManuel SchmuckManuel Le GalloGiovanni CherubiniLuca BeniniAbu SebastianAbbas RahimiNature PortfolioarticleScienceQENNature Communications, Vol 12, Iss 1, Pp 1-12 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Science Q
spellingShingle	Science Q Geethan Karunaratne Manuel Schmuck Manuel Le Gallo Giovanni Cherubini Luca Benini Abu Sebastian Abbas Rahimi Robust high-dimensional memory-augmented neural networks
description	The implementation of memory-augmented neural networks using conventional computer architectures is challenging due to a large number of read and write operations. Here, Karunaratne, Schmuck et al. propose an architecture that enables analog in-memory computing on high-dimensional vectors at accuracy matching 32-bit software equivalent.
format	article
author	Geethan Karunaratne Manuel Schmuck Manuel Le Gallo Giovanni Cherubini Luca Benini Abu Sebastian Abbas Rahimi
author_facet	Geethan Karunaratne Manuel Schmuck Manuel Le Gallo Giovanni Cherubini Luca Benini Abu Sebastian Abbas Rahimi
author_sort	Geethan Karunaratne
title	Robust high-dimensional memory-augmented neural networks
title_short	Robust high-dimensional memory-augmented neural networks
title_full	Robust high-dimensional memory-augmented neural networks
title_fullStr	Robust high-dimensional memory-augmented neural networks
title_full_unstemmed	Robust high-dimensional memory-augmented neural networks
title_sort	robust high-dimensional memory-augmented neural networks
publisher	Nature Portfolio
publishDate	2021
url	https://doaj.org/article/b553751b28d34003af781f5a72f0719a
work_keys_str_mv	AT geethankarunaratne robusthighdimensionalmemoryaugmentedneuralnetworks AT manuelschmuck robusthighdimensionalmemoryaugmentedneuralnetworks AT manuellegallo robusthighdimensionalmemoryaugmentedneuralnetworks AT giovannicherubini robusthighdimensionalmemoryaugmentedneuralnetworks AT lucabenini robusthighdimensionalmemoryaugmentedneuralnetworks AT abusebastian robusthighdimensionalmemoryaugmentedneuralnetworks AT abbasrahimi robusthighdimensionalmemoryaugmentedneuralnetworks
_version_	1718382881349304320

Robust high-dimensional memory-augmented neural networks

Ejemplares similares