A Discussion of ‘Adversarial Examples Are Not Bugs, They Are Features’: Adversarially Robust Neural Style Transfer

An experiment showing adversarial robustness makes neural style transfer work on a non-VGG architecture