Robustifying models against adversarial attacks by Langevin dynamics

Adversarial attacks on deep learning models have compromised their performance considerably. As remedies, a number of defense methods were proposed, which however, have been circumvented by newer and more sophisticated attacking strategies. In the midst of this ensuing arms race, the problem of robu...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	Neural networks Jg. 137; S. 1 - 17
Hauptverfasser:	Srinivasan, Vignesh, Rohrer, Csaba, Marban, Arturo, Müller, Klaus-Robert, Samek, Wojciech, Nakajima, Shinichi
Format:	Journal Article
Sprache:	Englisch
Veröffentlicht:	United States Elsevier Ltd 01.05.2021
Schlagworte:	Adversarial examples Langevin dynamics Robustness Adversarial examples Robustness Langevin dynamics
ISSN:	0893-6080, 1879-2782, 1879-2782
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!