Interpolated Adversarial Training: Achieving Robust Neural Networks Without Sacrificing Too Much Accuracy.

Alex Lamb, Vikas Verma, Juho Kannala, Yoshua Bengio

Browse the full CCS paper archive.