The Odds are Odd: A Statistical Test for Detecting Adversarial Examples

arXiv: Learning, 2019.

Cited by: 33|Views25
EI

Abstract:

We investigate conditions under which test statistics exist that can reliably detect examples, which have been adversarially manipulated in a white-box attack. These statistics can be easily computed and calibrated by randomly corrupting inputs. They exploit certain anomalies that adversarial attacks introduce, in particular if they follo...More

Code:

Data:

Your rating :
0

 

Tags
Comments