Question Answering 벤치마크
Question Answering on HotpotQA
JOINT-F1
- 2018-09-25 — SAFSR model: JOINT-F1 0.598
- 2019-05-16 — DFGN: JOINT-F1 0.5982
- 2019-11-09 — HGN + SemanticRetrievalMRS IR: JOINT-F1 0.599
- 2019-11-24 — Robustly Fine-tuned Graph-based Recurrent Retriever: JOINT-F1 0.612
- 2020-07-28 — BigBird-etc: JOINT-F1 0.736
- 2023-08-17 — Beam Retrieval: JOINT-F1 0.775
| Rank | Model | JOINT-F1 | ANS-EM | ANS-F1 | SUP-EM | SUP-F1 | JOINT-EM | Paper | Code | Year |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Beam Retrieval | 0.775 | 0.727 | 0.850 | 0.663 | 0.901 | 0.505 | End-to-End Beam Retrieval for Multi-Hop Question Answering | ShayekhBinIslam/openrag · canghongjian/beam_retriever · Alab-NII/2wikimultihop | 2023 |
| 2 | BigBird-etc | 0.736 | – | 0.755 | – | 0.891 | – | Big Bird: Transformers for Longer Sequences | huggingface/transformers · tensorflow/models · PaddlePaddle/PaddleNLP · +11 | 2020 |
| 3 | AISO | 0.720 | 0.675 | 0.805 | 0.612 | 0.860 | 0.449 | Adaptive Information Seeking for Open-Domain Question Answering | zycdev/aiso | 2021 |
| 4 | Chain-of-Skills | 0.717 | 0.674 | 0.801 | 0.613 | 0.853 | 0.457 | Chain-of-Skills: A Configurable Model for Open-domain Question Answering | mayer123/udt-qa | 2023 |
| 5 | TPRR | 0.708 | 0.670 | 0.795 | 0.594 | 0.843 | 0.444 | |||
| 6 | HopRetriever + Sp-search | 0.706 | 0.671 | 0.799 | 0.574 | 0.835 | 0.432 | HopRetriever: Retrieve Hops over Wikipedia to Answer Complex Questions | 2020 | |
| 7 | EBS-Large | 0.700 | 0.662 | 0.793 | 0.573 | 0.840 | 0.420 | |||
| 8 | HopRetriever | 0.698 | 0.671 | 0.799 | 0.572 | 0.826 | 0.431 | |||
| 9 | IRRR+ | 0.696 | 0.663 | 0.791 | 0.569 | 0.832 | 0.428 | Answering Open-Domain Questions of Varying Reasoning Steps from Text | beerqa/irrr | 2020 |
| 10 | EBS-SH | 0.689 | 0.655 | 0.786 | 0.559 | 0.831 | 0.409 | |||
| 11 | IRRR | 0.686 | 0.657 | 0.782 | 0.559 | 0.821 | 0.421 | Answering Open-Domain Questions of Varying Reasoning Steps from Text | beerqa/irrr | 2020 |
| 12 | HopRetriever-V2 | 0.678 | 0.648 | 0.778 | 0.561 | 0.818 | 0.410 | |||
| 13 | AFSGraph-retriever | 0.670 | 0.646 | 0.778 | 0.557 | 0.812 | 0.411 | |||
| 14 | Recursive Dense Retriever | 0.666 | 0.623 | 0.753 | 0.575 | 0.809 | 0.418 | Answering Complex Open-Domain Questions with Multi-Hop Dense Retrieval | facebookresearch/multihop_dense_retrieval | 2020 |
| 15 | Step-by-Step Retriever | 0.662 | 0.630 | 0.754 | 0.546 | 0.800 | 0.404 | |||
| 16 | DDRQA | 0.639 | 0.625 | 0.759 | 0.510 | 0.789 | 0.360 | Answering Any-hop Open-domain Questions with Iterative Document Reranking | 2020 | |
| 16 | HopRetriever-V1 | 0.639 | 0.608 | 0.739 | 0.531 | 0.793 | 0.380 | |||
| 18 | DR model large | 0.630 | 0.620 | 0.753 | 0.499 | 0.778 | 0.354 | |||
| 19 | Model name | 0.629 | 0.617 | 0.746 | 0.500 | 0.772 | 0.368 | |||
| 19 | HopAns | 0.629 | 0.617 | 0.746 | 0.500 | 0.772 | 0.368 |