Question Answering on TriviaQA Wikipedia (dev-full)
61.3F1MINIMAL (Our Selector)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| MINIMAL (Our Selector)n sent=20.4, Acc=95.3, Sp=x3.42018.05 | 61.3 | 55.6 | — | — | |
| MINIMAL (Our Selector)n sent=10.5, Acc=90.9, Sp=x6.62018.05 | 60.5 | 54.9 | — | — | |
| MINIMAL (Our Selector)n sent=10.5, Selector=Dyn2018.05 | 60.5 | 54.9 | 90.9 | 6.6 | |
| TF-IDFn sent=20, Acc=85.5, Sp=x3.52018.05 | 60.4 | 54.8 | — | — | |
| FULLn sent=69, Acc=95.9, Sp=x1.02018.05 | 59.6 | 53.5 | — | — | |
| FULLn sent=692018.05 | 59.6 | 53.5 | 95.9 | 1 | |
| MINIMAL (Our Selector)n sent=5.0, Acc=84.9, Sp=x13.82018.05 | 59.5 | 54 | — | — | |
| MINIMAL (Our Selector)n sent=5.0, Selector=Dyn2018.05 | 59.5 | 54 | 84.9 | 13.8 | |
| TF-IDFn sent=10, Acc=79.9, Sp=x6.92018.05 | 57.2 | 51.5 | — | — | |
| MINIMAL (TF-IDF)n sent=10, Selector=TF-IDF2018.05 | 57.2 | 51.5 | 79.9 | 6.9 | |
| Neural Cascades2018.05 | 56 | 51.6 | — | — | |
| Neural CascadesRank=12018.05 | 56 | 51.6 | — | — | |
| Reading Twiceevaluated_on=test set2018.05 | 55.1 | 48.6 | — | — | |
| Reading TwiceRank=22018.05 | 55.1 | 48.6 | — | — | |
| Mnemonic Readerevaluated_on=test set2018.05 | 52.9 | 46.9 | — | — | |
| Mnemonic ReaderRank=32018.05 | 52.9 | 46.9 | — | — | |
| TF-IDFn sent=5, Acc=73.0, Sp=x13.82018.05 | 51.9 | 45.8 | — | — | |
| MINIMAL (TF-IDF)n sent=5, Selector=TF-IDF2018.05 | 51.9 | 45.8 | 73 | 13.8 | |
| MEMEN2018.05 | 46.9 | 43.2 | — | — |