4 ms·
adversarial noise is very popular in the media but imo is a complete dead end for the desired goals - representations do not transfer between different models t
by whimsicalism 2y ago
adversarial noise is very popular in the media but imo is a complete dead end for the desired goals - representations do not transfer between different models this easily
- dijksterhuis 2y agoadversarial noise [transferability] for image classification used to be very easy (no idea now, not been in the space for half a decade). the [transferability] rates just drop off significantly for audio (always felt it was a similar vibe to RNN ‘vanishing gradients’) edit — specifically mention transferability