2 ms·
The whole point of GANs is to train a model until it can defeat detection by another model. Virtually any technique for detection that can be reproduced could j
by macksd 5y ago
The whole point of GANs is to train a model until it can defeat detection by another model. Virtually any technique for detection that can be reproduced could just be incorporated into the training loop and rendered almost useless.
- ford_o 5y agoHowever, the detection process could be so slow, that it wouldn't be practical.