2021 IJCNLP IJCNLP 2021

Better Robustness by More Coverage: Adversarial and Mixup Data Augmentation for Robust Finetuning