News for hackers. Stuff that matters.

A comment on Keras reimplementation of "One pixel attack for fooling deep neural networks". Parent comment

Re:Keras reimplementation of "One pixel attack for fooling d... by argonaut on Saturday February 24, 2018 @01:50AM (#16452118)
It also doesn't work for adversarial attacks. "Adversarial training" helps somewhat but is still very susceptible to adversarial attacks.
[ Parent | on HN ]

There are no comments on this comment yet.