Efficient Trigger Word Insertion

Yueqi Zeng, Ziqiang Li, Pengfei Xia +2

With the boom in the natural language processing (NLP) field these years, backdoor attacks pose immense threats against deep neural network models. However, previous works hardly c…