Are ConvNets better than humans on ImageNet? "Humans"on this plot = one human, who describes his experience here: http://karpathy.github.io/2014/09/02/what-i-learned-from-competing-against-a-convnet-on-imagenet/ …pic.twitter.com/iwz3OeV72e
-
No. Evolution has had plenty of time to work on it. https://en.wikipedia.org/wiki/Crypsis pic.twitter.com/AJmNK0HMMa
Camouflage != Adversarial Examples. AEs usually target one specific NN and don't work with other NNs. Proof: I still classify rhs as Panda.pic.twitter.com/QnPe5iysYJ
-
-
Again incorrect. There's plenty that's been done on black box adv attacks which succeed with high probability across nets. Such as in:
- Show replies
New conversation -
-
-
Anyway point is NatSelect+predprey+camo's exactly a search for perturbations of expressed pigmentation adversely targeting visual perception
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.