First they [adversarial examples] came for supervised learning...then RL...now variational autoencoders:https://openreview.net/forum?id=B1tExikAW¬eId=B1tExikAW …
Am I missing something or does this not make much sense? Directly flipping bits in the latent space isn’t an adversarial input.
-
-
The latent space is supposed to be a compressed representation. It’s good if a bitflip can change a lot in a meaningful way.
-
dunno, haven't read.
- 1 more reply
New conversation -
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.