Title: and the of Logit

Authors: Logan Engstrom, Andrew Ilyas, Anish Athalye

Abstract: We evaluate the robustness of Adversarial Logit Pairing, a recently proposed defense against adversarial examples. We find that a network trained with Adversarial Logit Pairing achieves 0.6% accuracy in the threat model in which the defense is considered. We provide a brief overview of the defense and the threat models/claims considered, as well as a discussion of the methodology and results of our attack, which may offer insights into the underlying the vulnerability of ALP to adversarial attack.

PDF link Landing page

Source link
thanks you RSS link
( https://www.reddit.com/r/MachineLearning/comments/92dmw3/r_evaluating_and_understanding_the_robustness_of/)


Please enter your comment!
Please enter your name here