http://www.reddit.com/user/MXXIV666
MXXIV6667 karma2022-11-03 17:54:38 UTC
Isn't it possible to directly calculate what should a "not a human" pattern look like instead of using an adversarial network?
Of course, this is for a case where you know the structure and weights of the network you want to fool.
View HistoryShare Link
MXXIV6667 karma
Isn't it possible to directly calculate what should a "not a human" pattern look like instead of using an adversarial network?
Of course, this is for a case where you know the structure and weights of the network you want to fool.
View HistoryShare Link