|
|
|
|
|
|
by nostrademons
3590 days ago
|
|
|
Markov chains aren't really relevant here - they're a way to generate random but "seemingly plausible" data - but yes, there is a way to fool the type of convolutional neural networks used in image recognition. Paper here: http://www.evolvingai.org/fooling |
|
|