Semantically Improved Adversarial Attack Based on Masked Language Model via Context Preservation.

Hao Tian, Hao-Tian Wu, Yiu-Ming Cheung, Junhui He, Zhihong Tian

Browse the full DSN paper archive.