Complexity control by gradient descent in deep networks

Understanding the underlying mechanisms behind the successes of deep networks remains a challenge. Here, the author demonstrates an implicit regularization in training deep networks, showing that the control of complexity in the training is hidden within the optimization technique of gradient descen...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Tomaso Poggio, Qianli Liao, Andrzej Banburski
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2020
Materias:	Science Q
Acceso en línea:	https://doaj.org/article/a50e459c37be4efd87800b05dd1d5ce3
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:a50e459c37be4efd87800b05dd1d5ce3
record_format	dspace
spelling	oai:doaj.org-article:a50e459c37be4efd87800b05dd1d5ce32021-12-02T14:28:23ZComplexity control by gradient descent in deep networks10.1038/s41467-020-14663-92041-1723https://doaj.org/article/a50e459c37be4efd87800b05dd1d5ce32020-02-01T00:00:00Zhttps://doi.org/10.1038/s41467-020-14663-9https://doaj.org/toc/2041-1723Understanding the underlying mechanisms behind the successes of deep networks remains a challenge. Here, the author demonstrates an implicit regularization in training deep networks, showing that the control of complexity in the training is hidden within the optimization technique of gradient descent.Tomaso PoggioQianli LiaoAndrzej BanburskiNature PortfolioarticleScienceQENNature Communications, Vol 11, Iss 1, Pp 1-5 (2020)
institution	DOAJ
collection	DOAJ
language	EN
topic	Science Q
spellingShingle	Science Q Tomaso Poggio Qianli Liao Andrzej Banburski Complexity control by gradient descent in deep networks
description	Understanding the underlying mechanisms behind the successes of deep networks remains a challenge. Here, the author demonstrates an implicit regularization in training deep networks, showing that the control of complexity in the training is hidden within the optimization technique of gradient descent.
format	article
author	Tomaso Poggio Qianli Liao Andrzej Banburski
author_facet	Tomaso Poggio Qianli Liao Andrzej Banburski
author_sort	Tomaso Poggio
title	Complexity control by gradient descent in deep networks
title_short	Complexity control by gradient descent in deep networks
title_full	Complexity control by gradient descent in deep networks
title_fullStr	Complexity control by gradient descent in deep networks
title_full_unstemmed	Complexity control by gradient descent in deep networks
title_sort	complexity control by gradient descent in deep networks
publisher	Nature Portfolio
publishDate	2020
url	https://doaj.org/article/a50e459c37be4efd87800b05dd1d5ce3
work_keys_str_mv	AT tomasopoggio complexitycontrolbygradientdescentindeepnetworks AT qianliliao complexitycontrolbygradientdescentindeepnetworks AT andrzejbanburski complexitycontrolbygradientdescentindeepnetworks
_version_	1718391271195672576

Complexity control by gradient descent in deep networks

Ejemplares similares