A Geometric Perspective on the Transferability of Adversarial Directions

State-of-the-art machine learning models frequently misclassify inputs that have been perturbed in an adversarial manner. Adversarial perturbations generated for a given input and a specific classifier often seem to be effective on other inputs and even different classifiers. In other words, adversa...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	arXiv.org 2018-11
Hauptverfasser:	Zachary, Charles, Rosenberg, Harrison, Papailiopoulos, Dimitris
Format:	Artikel
Sprache:	eng
Schlagworte:	Classifiers Data points Machine learning Neural networks Separators
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!