DeepFool: A Simple and Accurate Method to Fool Deep Neural Networks

State-of-the-art deep neural networks have achieved impressive results on many image classification tasks. However, these same architectures have been shown to be unstable to small, well sought, perturbations of the images. Despite the importance of this phenomenon, no effective methods have been pr...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) s. 2574 - 2582
Hlavní autori: Moosavi-Dezfooli, Seyed-Mohsen, Fawzi, Alhussein, Frossard, Pascal
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 01.06.2016
Predmet:
ISSN:1063-6919
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Abstract State-of-the-art deep neural networks have achieved impressive results on many image classification tasks. However, these same architectures have been shown to be unstable to small, well sought, perturbations of the images. Despite the importance of this phenomenon, no effective methods have been proposed to accurately compute the robustness of state-of-the-art deep classifiers to such perturbations on large-scale datasets. In this paper, we fill this gap and propose the DeepFool algorithm to efficiently compute perturbations that fool deep networks, and thus reliably quantify the robustness of these classifiers. Extensive experimental results show that our approach outperforms recent methods in the task of computing adversarial perturbations and making classifiers more robust.
AbstractList State-of-the-art deep neural networks have achieved impressive results on many image classification tasks. However, these same architectures have been shown to be unstable to small, well sought, perturbations of the images. Despite the importance of this phenomenon, no effective methods have been proposed to accurately compute the robustness of state-of-the-art deep classifiers to such perturbations on large-scale datasets. In this paper, we fill this gap and propose the DeepFool algorithm to efficiently compute perturbations that fool deep networks, and thus reliably quantify the robustness of these classifiers. Extensive experimental results show that our approach outperforms recent methods in the task of computing adversarial perturbations and making classifiers more robust.
Author Moosavi-Dezfooli, Seyed-Mohsen
Fawzi, Alhussein
Frossard, Pascal
Author_xml – sequence: 1
  givenname: Seyed-Mohsen
  surname: Moosavi-Dezfooli
  fullname: Moosavi-Dezfooli, Seyed-Mohsen
  email: seyed.moosavi@epfl.ch
  organization: Ecole Polytech. Fed. de Lausanne, Lausanne, Switzerland
– sequence: 2
  givenname: Alhussein
  surname: Fawzi
  fullname: Fawzi, Alhussein
  email: alhussein.fawzi@epfl.ch
  organization: Ecole Polytech. Fed. de Lausanne, Lausanne, Switzerland
– sequence: 3
  givenname: Pascal
  surname: Frossard
  fullname: Frossard, Pascal
  email: pascal.frossard@epfl.ch
  organization: Ecole Polytech. Fed. de Lausanne, Lausanne, Switzerland
BookMark eNotjslOwzAURQ0CiVKyZMXGP5Dwnu14YBeFFpDKIKZtZeIXEUjjKglC_D2pYHUW5-jqHrODLnbE2ClChgjuvHx9eMwEoM6EFXssccai0kZamyPusxmClql26I5YMgwfAIBOW7RuxspLou0yxvaCF_yp2Wxb4r4LvKiqr96PxG9pfI-Bj5HvKr7L-R1Nrp0wfsf-czhhh7VvB0r-OWcvy8VzeZ2u7q9uymKVNgLVmKJ3NVkHwSgt8-mfAe1NCBoQqreckCgnWStrzaSU1z4EWakgKidNELWcs7O_3YaI1tu-2fj-Z22MBZ2j_AWEY0qi
CODEN IEEPAD
ContentType Conference Proceeding
DBID 6IE
6IH
CBEJK
RIE
RIO
DOI 10.1109/CVPR.2016.282
DatabaseName IEEE Electronic Library (IEL) Conference Proceedings
IEEE Proceedings Order Plan (POP) 1998-present by volume
IEEE Xplore All Conference Proceedings
IEEE Electronic Library (IEL)
IEEE Proceedings Order Plans (POP) 1998-present
DatabaseTitleList
Database_xml – sequence: 1
  dbid: RIE
  name: IEEE/IET Electronic Library
  url: https://ieeexplore.ieee.org/
  sourceTypes: Publisher
DeliveryMethod fulltext_linktorsrc
Discipline Applied Sciences
Computer Science
EISBN 9781467388511
1467388513
EISSN 1063-6919
EndPage 2582
ExternalDocumentID 7780651
Genre orig-research
GroupedDBID 23M
29F
29O
6IE
6IH
6IK
ABDPE
ACGFS
ALMA_UNASSIGNED_HOLDINGS
CBEJK
IPLJI
M43
RIE
RIO
RNS
ID FETCH-LOGICAL-i214t-1a9fe890d74635467706a7dd6010cb5e1ee5e3f48877704a6add3c4d2c937d2f3
IEDL.DBID RIE
ISICitedReferencesCount 3332
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000400012302068&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
IngestDate Wed Aug 27 01:54:52 EDT 2025
IsDoiOpenAccess false
IsOpenAccess true
IsPeerReviewed false
IsScholarly true
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-i214t-1a9fe890d74635467706a7dd6010cb5e1ee5e3f48877704a6add3c4d2c937d2f3
OpenAccessLink https://infoscience.epfl.ch/handle/20.500.14299/125711
PageCount 9
ParticipantIDs ieee_primary_7780651
PublicationCentury 2000
PublicationDate 2016-June
PublicationDateYYYYMMDD 2016-06-01
PublicationDate_xml – month: 06
  year: 2016
  text: 2016-June
PublicationDecade 2010
PublicationTitle 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
PublicationTitleAbbrev CVPR
PublicationYear 2016
Publisher IEEE
Publisher_xml – name: IEEE
SSID ssj0001968189
ssj0023720
ssj0003211698
Score 2.6075015
Snippet State-of-the-art deep neural networks have achieved impressive results on many image classification tasks. However, these same architectures have been shown to...
SourceID ieee
SourceType Publisher
StartPage 2574
SubjectTerms Algorithm design and analysis
Computer vision
Level set
Neural networks
Optimization
Pattern recognition
Robustness
Title DeepFool: A Simple and Accurate Method to Fool Deep Neural Networks
URI https://ieeexplore.ieee.org/document/7780651
WOSCitedRecordID wos000400012302068&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV07b8IwELYo6tCpD6j6loeODSSOY-NuiBZ1aBHqS2zIsS8SEkoQhP7-nk1Ili6dkpw8RHey7853932E3DOe6swKFegoFQFnRgQ6kWGQcmkTCFPLMj8o_Conk8FspqYt8lDPwgCAbz6Dnnv1tXxbmK27KutL6cqAmOscSCl3s1rNfYoS6HtU_R1jZiNUXVFgjo2lwdjsj76n766xS_SYx-BrmFW8Yxkf_--XTki3mdCj09r3nJIW5GfkuAopabVhNyjaszbsZR0yegJYjYti-UiH9GPh0IGpzi0dGrN1uBH0zZNK07KgbhV1y6nD8NBLfPim8U2XfI2fP0cvQUWlECxYxMsg0iqDgQqt5Bhh4OEoQ6GltS4dM2kCEUACcYa7GTUbci3w2IsNt8xg-IL2is9JOy9yuCBUafRnWWIza9CgaYb5mpHcsBRisJg8XZKOU9N8tUPLmFcauvpbfE2OnBV2zVc3pF2ut3BLDs1Pudis77yJfwHuXKSN
linkProvider IEEE
linkToHtml http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3PT8IwFG6ImugJFYy_7cGjg63rWuqNoAQjEKJouJGufUtIyEZg-PfblrFdvHja9trD0pf2vdf33vch9EhoLBPNhCeDmHmUKObJiPteTLmOwI81SVyj8JCPx53ZTExq6KnshQEAV3wGLfvqcvk6U1t7Vdbm3KYBTaxzGFFKgl23VnWjIpixPqL8Dk1sw0SZUyCWj6VC2Wz3vicftrSLtYhD4au4VZxp6df_91OnqFn16OFJaX3OUA3Sc1QvnEpcbNmNEe15G_ayBuq9AKz6WbZ8xl38ubD4wFimGneV2lrkCDxytNI4z7Cdhe10bFE85NI8XNn4pom--q_T3sAryBS8BQlo7gVSJNARvubU-BjmeOQ-k1xrG5CpOIIAIIIwMfuZmyEqmTn4QkU1UcaBMRoLL9BBmqVwibCQxqIlkU60MiqNExOxKU4ViSEEbcKnK9SwyzRf7fAy5sUKXf8tfkDHg-loOB--jd9v0InVyK4U6xYd5Ost3KEj9ZMvNut7p-5fsCCn1A
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Abook&rft.genre=proceeding&rft.title=2016+IEEE+Conference+on+Computer+Vision+and+Pattern+Recognition+%28CVPR%29&rft.atitle=DeepFool%3A+A+Simple+and+Accurate+Method+to+Fool+Deep+Neural+Networks&rft.au=Moosavi-Dezfooli%2C+Seyed-Mohsen&rft.au=Fawzi%2C+Alhussein&rft.au=Frossard%2C+Pascal&rft.date=2016-06-01&rft.pub=IEEE&rft.eissn=1063-6919&rft.spage=2574&rft.epage=2582&rft_id=info:doi/10.1109%2FCVPR.2016.282&rft.externalDocID=7780651