thunlp / TAADpapers

Must-read Papers on Textual Adversarial Attack and Defense
MIT License
1.49k stars 194 forks source link

Add a NeurIPS 2020 paper #24

Closed shizhouxing closed 3 years ago

shizhouxing commented 3 years ago

This is a general framework for robustness verification and certified training. There are also experiments on training text classifiers that are certifiably robust to word substitutions.

Fanchao-Qi commented 3 years ago

Thanks for your contribution.