Question Rewriting on CANARD (test)
0.4133Hard Difficulty ScoreP-hard
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| P-hardModel Architecture=BART-base, Training Strategy=Adapter Tuning2023.11 | 0.4133 | 0.4639 | 0.5524 | 0.4766 | |
| Mix-goldModel Architecture=BART-base, Training Strategy=Adapter Tuning2023.11 | 0.4133 | 0.5468 | 0.7363 | 0.5655 | |
| Mix-goldModel Architecture=BART-base, Training Strategy=Fine-tuning2023.11 | 0.4091 | 0.5615 | 0.74 | 0.5702 | |
| SModel Architecture=BART-base, Training Strategy=Fine-tuning2023.11 | 0.3938 | 0.537 | 0.6633 | 0.5314 | |
| SModel Architecture=BART-base, Training Strategy=Adapter Tuning2023.11 | 0.392 | 0.5314 | 0.6597 | 0.5277 | |
| SLADModel Architecture=BART-base, Training Strategy=Adapter Tuning2023.11 | 0.3826 | 0.5422 | 0.6757 | 0.5335 | |
| SLAFModel Architecture=BART-base, Training Strategy=Adapter Tuning2023.11 | 0.3455 | 0.5605 | 0.6905 | 0.5322 | |
| P-mediumModel Architecture=BART-base, Training Strategy=Adapter Tuning2023.11 | 0.3441 | 0.5468 | 0.6298 | 0.5069 | |
| SLAF-uni.Model Architecture=BART-base, Training Strategy=Adapter Tuning, Fusion Strategy=Uniform2023.11 | 0.3405 | 0.5588 | 0.6727 | 0.524 | |
| SLAF-predModel Architecture=BART-base, Training Strategy=Adapter Tuning, Fusion Strategy=Predicted2023.11 | 0.3296 | 0.5562 | 0.7083 | 0.5314 | |
| Mix-goldModel Architecture=LSTM-based2023.11 | 0.2779 | 0.5191 | 0.8653 | 0.5541 | |
| P-easyModel Architecture=BART-base, Training Strategy=Adapter Tuning2023.11 | 0.2742 | 0.5555 | 0.7363 | 0.522 | |
| SModel Architecture=LSTM-based2023.11 | 0.2629 | 0.5079 | 0.7941 | 0.5216 |