search

thunlp / TAADpapers

Must-read Papers on Textual Adversarial Attack and Defense
MIT License
1.51k stars 195 forks source link

Update README.md revise some errors && add code url #50

Closed rick-lzr closed 7 months ago

rick-lzr commented 7 months ago

1、In the paper ’Adversarial Text Generation by Search and Learning‘ , equation 2 uses the confidence scores output of the model to calculate word importance. I believe this should be a score-based method rather than a decision-based method. image

2、add 2 code url Bridge the Gap Between CV and NLP! A Gradient-based Textual Adversarial Attack Framework. ->https://github.com/Phantivia/T-PGD Adversarial Text Generation by Search and Learning. ->https://github.com/DABAI6666/ATGSL

yangalan123 commented 7 months ago

Thanks for catching this problem! Yeah, your correction looks reasonable. Also thanks for adding the code link to the paper!