Evasion attack of multi-class linear classifiers

Han Xiao, Thomas Stibor, Claudia Eckert

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Scopus citations

Abstract

Machine learning has yield significant advances in decision-making for complex systems, but are they robust against adversarial attacks? We generalize the evasion attack problem to the multi-class linear classifiers, and present an efficient algorithm for approximating the optimal disguised instance. Experiments on real-world data demonstrate the effectiveness of our method.

Original languageEnglish
Title of host publicationAdvances in Knowledge Discovery and Data Mining - 16th Pacific-Asia Conference, PAKDD 2012, Proceedings
Pages207-218
Number of pages12
EditionPART 1
DOIs
StatePublished - 2012
Event16th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2012 - Kuala Lumpur, Malaysia
Duration: 29 May 20121 Jun 2012

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 1
Volume7301 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference16th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2012
Country/TerritoryMalaysia
CityKuala Lumpur
Period29/05/121/06/12

Fingerprint

Dive into the research topics of 'Evasion attack of multi-class linear classifiers'. Together they form a unique fingerprint.

Cite this