Question Answering on SQuAD-Open
59.5EMAISO_large
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| AISO_large# read=24.8, Pretrained Model Scale=Large2021.09 | 59.5 | 67.6 | |
| SPARTAPretrained Model Scale=Large2021.09 | 59.3 | 66.5 | |
| IRRR# read=>= 150, Pretrained Model Scale=Large2021.09 | 56.8 | 63.2 | |
| Fusion-in-DecoderModel scale=large2020.07 | 56.7 | 63.2 | |
| Path Retriever2020.07 | 56.5 | 63.8 | |
| GRR# read=>= 500, Pretrained Model Scale=Large2021.09 | 56.5 | 63.8 | |
| Fusion-in-DecoderModel scale=base2020.07 | 53.4 | 60.6 | |
| Multi-Passage BERT2020.07 | 53 | 60.9 | |
| Multi-passage BERT# read=1002021.09 | 53 | 60.9 | |
| MUPPETRepresentation level=sentence-level2019.06 | 39.3 | 46.2 | |
| MUPPET# read=452021.09 | 39.3 | 46.2 | |
| BERTserini2019.06 | 38.6 | 46.1 | |
| DPR2020.07 | 36.7 | — | |
| BM25+DPR# read=1002021.09 | 36.7 | — | |
| MUPPETRepresentation level=paragraph-level2019.06 | 35.6 | 42.5 | |
| Minimal2019.06 | 34.7 | 42.6 | |
| TF-IDF + Reader2019.06 | 34.6 | 41.6 | |
| Multi-step2019.06 | 31.9 | 39.2 | |
| Multi-step Reasoner# read=52021.09 | 31.9 | 39.2 | |
| Par. Ranker + Full Agg.2019.06 | 30.2 | — | |
| DrQAmultitask=true2019.06 | 29.8 | — | |
| DrQA2020.07 | 29.8 | — | |
| DPR# read=1002021.09 | 29.8 | — | |
| R32019.06 | 29.1 | 37.5 | |
| DS-QA2019.06 | 28.7 | 36.6 | |
| DrQAmultitask=false2019.06 | 28.4 | — | |
| DrQA# read=52021.09 | 27.1 | — | |
| ORQA2020.07 | 20.2 | — |