Evasion attack of multi-class linear classifiers

Han Xiao, Thomas Stibor, Claudia Eckert

Publikation: Beitrag in Buch/Bericht/KonferenzbandKonferenzbeitragBegutachtung

4 Zitate (Scopus)

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.

OriginalspracheEnglisch
TitelAdvances in Knowledge Discovery and Data Mining - 16th Pacific-Asia Conference, PAKDD 2012, Proceedings
Seiten207-218
Seitenumfang12
AuflagePART 1
DOIs
PublikationsstatusVeröffentlicht - 2012
Veranstaltung16th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2012 - Kuala Lumpur, Malaysia
Dauer: 29 Mai 20121 Juni 2012

Publikationsreihe

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NummerPART 1
Band7301 LNAI
ISSN (Print)0302-9743
ISSN (elektronisch)1611-3349

Konferenz

Konferenz16th Pacific-Asia Conference on Advances in Knowledge Discovery and Data Mining, PAKDD 2012
Land/GebietMalaysia
OrtKuala Lumpur
Zeitraum29/05/121/06/12

Fingerprint

Untersuchen Sie die Forschungsthemen von „Evasion attack of multi-class linear classifiers“. Zusammen bilden sie einen einzigartigen Fingerprint.

Dieses zitieren