The Blog



Share

Denoised smoothing: Provably defending pretrained classifiers against adversarial examples