Making Machine Learning Robust Against Adversarial Inputs.
The article discusses the failure of machine learning algorithms that exceed human performance in naturally occurring scenarios when an adversary is able to modify their input data. Topics include the use of model training, input validation and architectural changes to defend against adversaries, an...
| Published in: | Communications of the ACM Vol. 61; no. 7; pp. 56 - 67 |
|---|---|
| Main Authors: | , , |
| Format: | Article |
| Published: |
Association for Computing Machinery
Jul2018
|
| Subjects: | |
| Online Access: | View this record in EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=hlh&AN=130478724&site=ehost-live header: @attributes: shortDbName: hlh uiTerm: 130478724 longDbName: Humanities International Complete uiTag: AN controlInfo: bkinfo: jinfo: jid: 00010782 ACM jtl: Communications of the ACM issn: 00010782 maglogo: N pubinfo: dt: Jul2018 vid: 61 iid: 7 pid: 68 pub: Association for Computing Machinery artinfo: ui: 130478724 10.1145/3134599 ppf: 56 ppct: 11 formats: tig: atl: Making Machine Learning Robust Against Adversarial Inputs. aug: au: GOODFELLOW, IAN MCDANIEL, PATRICK PAPERNOT, NICOLAS affil: Staff research scientist at Google Brain, Mountain View, CA, USA Inventor of Generative Adversarial Networks William L. Weiss Professor of Information and Communications Technology in the School of Electrical Engineering and Computer Science at Pennsylvania State University, University Park, PA, USA Fellow of IEEE Fellow of ACM Google Ph.D. Fellow in Security in the Department of Computer Science and Engineering at Penn State University, University Park, PA, USA su: Machine learning Algorithms Computer input design Computer programming Computer security Security systems sug: subj: Machine learning Algorithms Computer input design Computer programming Computer security Security systems ab: The article discusses the failure of machine learning algorithms that exceed human performance in naturally occurring scenarios when an adversary is able to modify their input data. Topics include the use of model training, input validation and architectural changes to defend against adversaries, and the need for more tools to verify machine learning models in order to prevent attacks by adversaries. pubtype: Periodical doctype: Article src: R language: English refInfo: copyright: @attributes: flag: Y dt: @attributes: year: 2018 holdings: @attributes: islocal: N |
|---|