New change on loss function about fakebob

[FAKEBOB-adversarial-attack]

Hello, thanks for your feedback. We have changed the early stop condition to fit the updated loss function.

BTW, the reason why we remove from the loss function the outermost maximum operation is that it will cause unexpected issue for benign voices whose initial loss is slightly larger than adver_thresh.

For example, suppose adver_thresh=0, and the initial loss of a benign voice is 0.01. This voice is not robust enough, so add a random noise is enough to make it adversarial. Hence, the samples_per_draw noisy samples are adversarial voices. If the outermost maximum operation is retained, all the loss of these noisy samples are zero. So the estimated gradient remains zero, and the sample will not be updated throughout the iteration procedure.

That 's why we can only achieve 99% ASR, and not 100% in our paper. Using the updated loss function, we can achieve 100% ASR.

[Li-Chi-Chang]

Thanks for sharing!