Humans can decipher adversarial images

Convolutional Neural Networks (CNNs) have reached human-level benchmarks in classifying images, but they can be “fooled” by adversarial examples that elicit bizarre misclassifications from machines. Here, the authors show how humans can anticipate which objects CNNs will see in adversarial images....

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Zhenglong Zhou, Chaz Firestone
Formato:	article
Lenguaje:	EN
Publicado:	Nature Portfolio 2019
Materias:	Science Q
Acceso en línea:	https://doaj.org/article/11bc950f138d40c9b8eaad6a445e6db4
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:11bc950f138d40c9b8eaad6a445e6db4
record_format	dspace
spelling	oai:doaj.org-article:11bc950f138d40c9b8eaad6a445e6db42021-12-02T14:38:44ZHumans can decipher adversarial images10.1038/s41467-019-08931-62041-1723https://doaj.org/article/11bc950f138d40c9b8eaad6a445e6db42019-03-01T00:00:00Zhttps://doi.org/10.1038/s41467-019-08931-6https://doaj.org/toc/2041-1723Convolutional Neural Networks (CNNs) have reached human-level benchmarks in classifying images, but they can be “fooled” by adversarial examples that elicit bizarre misclassifications from machines. Here, the authors show how humans can anticipate which objects CNNs will see in adversarial images.Zhenglong ZhouChaz FirestoneNature PortfolioarticleScienceQENNature Communications, Vol 10, Iss 1, Pp 1-9 (2019)
institution	DOAJ
collection	DOAJ
language	EN
topic	Science Q
spellingShingle	Science Q Zhenglong Zhou Chaz Firestone Humans can decipher adversarial images
description	Convolutional Neural Networks (CNNs) have reached human-level benchmarks in classifying images, but they can be “fooled” by adversarial examples that elicit bizarre misclassifications from machines. Here, the authors show how humans can anticipate which objects CNNs will see in adversarial images.
format	article
author	Zhenglong Zhou Chaz Firestone
author_facet	Zhenglong Zhou Chaz Firestone
author_sort	Zhenglong Zhou
title	Humans can decipher adversarial images
title_short	Humans can decipher adversarial images
title_full	Humans can decipher adversarial images
title_fullStr	Humans can decipher adversarial images
title_full_unstemmed	Humans can decipher adversarial images
title_sort	humans can decipher adversarial images
publisher	Nature Portfolio
publishDate	2019
url	https://doaj.org/article/11bc950f138d40c9b8eaad6a445e6db4
work_keys_str_mv	AT zhenglongzhou humanscandecipheradversarialimages AT chazfirestone humanscandecipheradversarialimages
_version_	1718390875370815488

Humans can decipher adversarial images

Ejemplares similares