Generative Adversarial Training with Perturbed Token Detection for Model Robustness.

Jiahao Zhao, Wenji Mao

Browse the full EMNLP paper archive.