Adversarial Text Detection on IMDB (test)
97.8F1 ScoreVoteTRANS
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| VoteTRANSTarget Model=DistilBERT, Attack=TextFooler, Auxiliary attack=Same2023.06 | 97.8 | 99.6 | |
| VoteTRANSTarget Model=DistilBERT, Attack=TextFooler, Auxiliary attack=PWWS2023.06 | 97.7 | 100 | |
| VoteTRANSTarget Model=DistilBERT, Attack=BAE, Auxiliary attack=Same2023.06 | 97.7 | 100 | |
| VoteTRANSTarget Model=BERT, Attack=PWWS, Auxiliary attack=Same2023.06 | 97.4 | 98.4 | |
| VoteTRANSTarget Model=DistilBERT, Attack=PWWS, Auxiliary attack=Same2023.06 | 96.9 | 98.4 | |
| VoteTRANSTarget Model=DistilBERT, Attack=BAE, Auxiliary attack=PWWS2023.06 | 96.3 | 99.2 | |
| VoteTRANSTarget Model=DistilBERT, Attack=IGA, Auxiliary attack=PWWS2023.06 | 95.9 | 99.2 | |
| VoteTRANSTarget Model=DistilBERT, Attack=IGA, Auxiliary attack=Same2023.06 | 95.8 | 99.2 | |
| VoteTRANSTarget Model=CNN, Attack=PWWS, Auxiliary attack=Same2023.06 | 95.4 | 98.8 | |
| VoteTRANSTarget Model=LSTM, Attack=PWWS, Auxiliary attack=Same2023.06 | 94.8 | 97.6 | |
| WDRTarget Model=DistilBERT, Attack=TextFooler, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 94.2 | 97.3 | |
| WDRTarget Model=BERT, Attack=PWWS, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 92.4 | 92.5 | |
| WDRTarget Model=DistilBERT, Attack=PWWS, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 92.1 | 94.2 | |
| FGWSTarget Model=BERT, Attack=PWWS, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 89.8 | 82.7 | |
| FGWSTarget Model=DistilBERT, Attack=PWWS, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 89.5 | 82.7 | |
| WDRTarget Model=DistilBERT, Attack=IGA, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 88.5 | 95.5 | |
| WDRTarget Model=DistilBERT, Attack=BAE, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 88 | 96.3 | |
| FGWSTarget Model=CNN, Attack=PWWS, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 86.3 | 79.6 | |
| FGWSTarget Model=DistilBERT, Attack=TextFooler, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 86 | 77.6 | |
| WDRTarget Model=CNN, Attack=PWWS, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 84.3 | 90 | |
| WDRTarget Model=LSTM, Attack=PWWS, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 84.1 | 86.8 | |
| FGWSTarget Model=DistilBERT, Attack=IGA, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 83.8 | 74.8 | |
| FGWSTarget Model=LSTM, Attack=PWWS, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 80 | 69.6 | |
| FGWSTarget Model=DistilBERT, Attack=BAE, Detector training configuration=DistilBERT, IMDB, PWWS2023.06 | 65.6 | 50.2 |