Research Article
Textual Backdoor Attack for the Text Classification System
Figure 11
Accuracy of target model on original sentences from the MR dataset and attack success rate of backdoor samples with the trigger at the beginning of the sentence according to the proportion of backdoor samples in the input dataset. Each pair of bars represents the performance of the target model trained using a different proportion of backdoor samples. (a) The trigger text is “Wow.” (b) The trigger text is “In other words.”
(a) |
(b) |