Towards Trustworthy NLP: An Adversarial Robustness Enhancement Based on Perplexity Difference.

Zhaocheng Ge, Hanping Hu, Tengfei Zhao

Browse the full ECAI paper archive.