Supporting Fact Prediction on HotpotQA full wiki (dev)
86.5F1 ScoreAISO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| AISOPre-trained Model Scale=large2021.09 | 86.5 | 61.5 | |
| TPRRPre-trained Model Scale=large, Source=Official Leaderboard2021.09 | 84.5 | 60.2 | |
| AISOPre-trained Model Scale=base2021.09 | 81.9 | 55.1 | |
| HopRetriever-plusVariant=+2021.09 | 81.8 | 56 | |
| MDRVariant=+2021.09 | 79.4 | 56.5 | |
| DDRQA2021.09 | 79.1 | 51.3 | |
| HopRetriever2021.09 | 78.9 | 52.5 | |
| Recurrent Graph-based RetrievalReader=BERT wwm2019.11 | 76.1 | 49.3 | |
| GRRVariant=+2021.09 | 76.1 | 49.2 | |
| Recurrent Graph-based RetrievalReader=BERT base2019.11 | 75 | 47.9 | |
| Transformer-XH2019.11 | 71.6 | 42.2 | |
| Semantic Retrieval2019.11 | 71.5 | 39.9 | |
| Semantic Retrieval2021.09 | 71.5 | 39.9 | |
| GoldEn Retriever2019.11 | 64.6 | — | |
| Cognitive Graph2019.11 | 58.5 | 23.1 | |
| CogQA2021.09 | 58.5 | 23.1 | |
| MUPPET2019.11 | 47.7 | 17 | |
| MUPPET2021.09 | 47.7 | 17 | |
| Baseline2019.11 | 41 | 5.3 |