Efficiently generating sentence-level textual adversarial examples with Seq2seq Stacked Auto-Encoder

Ang Li, Fangyuan Zhang, Shuangjiao Li, Tianhua Chen, Pan Su, Hongtao Wang

Abstract: Highlights•A novel and efficient sentence-level adversarial attack method is proposed.•A stacked auto-encoder network is devised to achieve naturality and similarity.•Experiments show promising performance in attack success rate and efficiency.