Show simple item record

dc.contributor.authorYang, Zhuo
dc.contributor.authorHan, Yufei
dc.contributor.authorZhang, Xiangliang
dc.date.accessioned2021-09-20T06:23:43Z
dc.date.available2020-12-20T13:52:07Z
dc.date.available2021-09-20T06:23:43Z
dc.date.issued2021
dc.identifier.issn2374-3468
dc.identifier.issn2159-5399
dc.identifier.urihttp://hdl.handle.net/10754/666523
dc.description.abstractEvasion attack in multi-label learning systems is an interesting, widely witnessed, yet rarely explored research topic. Characterizing the crucial factors determining the attackability of the multi-label adversarial threat is the key to interpret the origin of the adversarial vulnerability and to understand how to mitigate it. Our study is inspired by the theory of adversarial risk bound. We associate the attackability of a targeted multi-label classifier with the regularity of the classifier and the training data distribution. Beyond the theoretical attackability analysis, we further propose an efficient empirical attackability estimator via greedy label space exploration. It provides provably computational efficiency and approximation accuracy. Substantial experimental results on real-world datasets validate the unveiled attackability factors and the effectiveness of the proposed empirical attackability indicator.
dc.publisherarXiven_US
dc.relation.urlhttps://arxiv.org/pdf/2012.09427en_US
dc.rightsArchived with thanks to arXiv
dc.titleCharacterizing the Evasion Attackability of Multi-label Classifiers
dc.typeProceedings Paper
dc.contributor.departmentComputer Science Program
dc.contributor.departmentComputer, Electrical and Mathematical Science and Engineering (CEMSE) Division
dc.conference.dateFEB 02-09, 2021
dc.conference.name35th AAAI Conference on Artificial Intelligence / 33rd Conference on Innovative Applications of Artificial Intelligence / 11th Symposium on Educational Advances in Artificial Intelligence
dc.conference.locationELECTR NETWORK
dc.identifier.wosutWOS:000681269802038
dc.eprint.versionPost-print
dc.contributor.institutionNorton Research Group, Sophia Antipolis, France.en_US
dc.identifier.volume35
dc.identifier.pages10647-10655
dc.identifier.arxivid2012.09427
kaust.personYang, Zhuo
kaust.personZhang, Xiangliang
refterms.dateFOA2020-12-20T13:52:53Z


Files in this item

Thumbnail
Name:
Preprintfile1.pdf
Size:
5.362Mb
Format:
PDF
Description:
Pre-print

This item appears in the following Collection(s)

Show simple item record

VersionItemEditorDateSummary

*Selected version