Paul, S.; Chen, P.-Y. Vision Transformers Are Robust Learners. AAAI 2022, 36, 2071-2081.