Question Selection on Fairytale QA (test)
97.5Grammatical CorrectnessMa
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| MaDescription=Baseline a2022.09 | 97.5 | 100 | 95.8 | 94.3 | 92 | 92.2 | 90.5 | 94.6 | 87 | |
| MsDescription=Baseline s2022.09 | 96.4 | 98.8 | 94.4 | 95.5 | 92.5 | 93.4 | 91.2 | 94.6 | 87.5 | |
| bi-gramStrategy=n-gram similarity, Order=bi-gram2022.09 | 95.3 | 99.3 | 94.3 | 94.3 | 93.2 | 93.7 | 90.2 | 94.3 | 85.7 | |
| APSStrategy=APS2022.09 | 95.2 | 98.5 | 95.7 | 97.2 | 92.2 | 97.7 | 94.8 | 95.9 | 89.5 | |
| tri-gram + APSStrategy=tri-gram with APS2022.09 | 94.8 | 97.8 | 95.2 | 93.8 | 91.8 | 94 | 91.3 | 94.1 | 86.3 | |
| GTDescription=Ground Truth2022.09 | 94.5 | 96.3 | 94.2 | 93.7 | 88.5 | 92.8 | 89.2 | 92.7 | 86.7 | |
| bi-gram + APSStrategy=bi-gram with APS2022.09 | 94.5 | 98.8 | 96.2 | 94.2 | 92.5 | 94.5 | 93.2 | 94.8 | 88.8 | |
| bi-gram + APS + RTStrategy=bi-gram with APS and Round-trip2022.09 | 94.5 | 98.8 | 96 | 94.2 | 91.3 | 93.7 | 93.2 | 94.5 | 88.3 | |
| tri-gramStrategy=n-gram similarity, Order=tri-gram2022.09 | 94.3 | 98 | 92.2 | 93 | 90.5 | 92.7 | 85.8 | 92.4 | 83.8 | |
| bi-gram + RTStrategy=bi-gram with Round-trip2022.09 | 94.3 | 99.2 | 95.2 | 93.8 | 91 | 92.7 | 91.8 | 94 | 86 | |
| round-trip (RT)Strategy=Round-trip validation2022.09 | 92.8 | 97 | 94.5 | 93.7 | 88.8 | 93.5 | 87.8 | 92.6 | 86.2 | |
| APS + RTStrategy=APS with Round-trip2022.09 | 92.7 | 98.3 | 95.5 | 97.3 | 91.5 | 94.8 | 92.8 | 94.7 | 89.3 | |
| tri-gram + RTStrategy=tri-gram with Round-trip2022.09 | 92.7 | 97.3 | 93.2 | 92.8 | 89 | 93.5 | 88.2 | 92.4 | 85.7 | |
| tri-gram + APS + RTStrategy=tri-gram with APS and Round-trip2022.09 | 92.5 | 98.7 | 93.8 | 96.2 | 91 | 94.2 | 92.2 | 94.1 | 88.2 |