SSMBA: Self-Supervised Manifold Based Data Augmentation for Improving Out-of-Domain Robustness.

Nathan Ng, Kyunghyun Cho, Marzyeh Ghassemi

Browse the full EMNLP paper archive.