Adversarial Text Detection on IMDB
97.7F1 ScoreVoteTRANS_same
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| VoteTRANS_sameTarget Model=DistilBERT, Attack=BAE, Auxiliary Attack=BAE2023.06 | 97.7 | 100 | — | |
| VoteTRANS_sameTarget Model=DistilBERT, Attack=PWWS, Auxiliary Attack=PWWS2023.06 | 96.9 | 98.4 | — | |
| VoteTRANS_diffTarget Model=DistilBERT, Attack=BAE, Auxiliary Attack=PWWS2023.06 | 96.3 | 99.2 | — | |
| WDRModel=DistilBERT, Attack=TextFooler2022.04 | 94.2 | — | 97.3 | |
| WDRModel=BERT, Attack=PWWS2022.04 | 92.4 | — | 92.5 | |
| WDRModel=DistilBERT, Attack=PWWS2022.04 | 92.1 | — | 94.2 | |
| WDRTarget Model=DistilBERT, Attack=PWWS, Detector training configuration=(DistilBERT, AG News, PWWS)2023.06 | 91.4 | 93 | — | |
| WDRModel=DistilBERT, Attack=PWWS, Detector Trained On=DistilBERT, AG News, PWWS2022.04 | 91.4 | 93 | — | |
| FGWSModel=BERT, Attack=PWWS2022.04 | 89.8 | — | 82.7 | |
| FGWSTarget Model=DistilBERT, Attack=PWWS, Detector training configuration=(DistilBERT, AG News, PWWS)2023.06 | 89.5 | 82.7 | — | |
| FGWSModel=DistilBERT, Attack=PWWS, Detector Trained On=DistilBERT, AG News, PWWS2022.04 | 89.5 | 82.7 | — | |
| FGWSModel=DistilBERT, Attack=PWWS2022.04 | 89.5 | — | 82.7 | |
| WDRModel=DistilBERT, Attack=IGA2022.04 | 88.5 | — | 95.5 | |
| WDRTarget Model=DistilBERT, Attack=BAE, Detector training configuration=(DistilBERT, AG News, PWWS)2023.06 | 88.1 | 97 | — | |
| WDRModel=DistilBERT, Attack=BAE, Detector Trained On=DistilBERT, AG News, PWWS2022.04 | 88.1 | 97 | — | |
| WDRModel=DistilBERT, Attack=BAE2022.04 | 88 | — | 96.3 | |
| FGWSModel=CNN, Attack=PWWS2022.04 | 86.3 | — | 79.6 | |
| FGWSModel=DistilBERT, Attack=TextFooler2022.04 | 86 | — | 77.6 | |
| WDRModel=CNN, Attack=PWWS2022.04 | 84.3 | — | 90 | |
| WDRModel=LSTM, Attack=PWWS2022.04 | 84.1 | — | 86.8 | |
| FGWSModel=DistilBERT, Attack=IGA2022.04 | 83.8 | — | 74.8 | |
| FGWSModel=LSTM, Attack=PWWS2022.04 | 80 | — | 69.6 | |
| FGWSTarget Model=DistilBERT, Attack=BAE, Detector training configuration=(DistilBERT, AG News, PWWS)2023.06 | 65.6 | 55.2 | — | |
| FGWSModel=DistilBERT, Attack=BAE, Detector Trained On=DistilBERT, AG News, PWWS2022.04 | 65.6 | 55.2 | — | |
| FGWSModel=DistilBERT, Attack=BAE2022.04 | 65.6 | — | 50.2 |