HotpotQA

A Dataset for Diverse, Explainable Multi-hop Question Answering

What is HotpotQA?

HotpotQA is a question answering dataset featuring natural, multi-hop questions, with strong supervision for supporting facts to enable more explainable question answering systems. It is collected by a team of NLP researchers at Carnegie Mellon University, Stanford University, and Université de Montréal.

For more details about HotpotQA, please refer to our EMNLP 2018 paper:

Getting started

HotpotQA is distributed under a CC BY-SA 4.0 License. The training and development sets can be downloaded below.

A more comprehensive summary about data download, preprocessing, baseline model training, and evaluation is included in our GitHub repository, and linked below.

Once you have built your model, you can use the evaluation script we provide below to evaluate model performance by running python hotpot_evaluate_v1.py <path_to_prediction> <path_to_gold>

To submit your models and evaluate them on the official test sets, please read our submission guide hosted on Codalab.

We also release the processed Wikipedia used in the process of creating HotpotQA (also under a CC BY-SA 4.0 License), serving both as the corpus for the fullwiki setting in our evaluation, and hopefully as a standalone resource for future researches involving processed text on Wikipedia. Below please find the link to the documentation for this corpus.

Stay connected!

Join our Google group to receive updates or initiate discussions about HotpotQA!

If you use HotpotQA in your research, please cite our paper with the following BibTeX entry

@inproceedings{yang2018hotpotqa,
  title={{HotpotQA}: A Dataset for Diverse, Explainable Multi-hop Question Answering},
  author={Yang, Zhilin and Qi, Peng and Zhang, Saizheng and Bengio, Yoshua and Cohen, William W. and Salakhutdinov, Ruslan and Manning, Christopher D.},
  booktitle={Conference on Empirical Methods in Natural Language Processing ({EMNLP})},
  year={2018}
}
Leaderboard (Distractor Setting)
In the distractor setting, a question-answering system reads 10 paragraphs to provide an answer (Ans) to a question. They must also justify these answers with supporting facts (Sup).
Model Code Ans Sup Joint
EM F1 EM F1 EM F1
1
Dec 1, 2019		 HGN-large (single model)
Anonymous
69.22 82.19 62.76 88.47 47.11 74.21
2
Oct 18, 2019		 C2F Reader (single model)
Joint Laboratory of HIT and iFLYTEK Research
67.98 81.24 60.81 87.63 44.67 72.73
3
Nov 19, 2019		 SAE-large (single model)
JD AI Research
Tu, Huang et al., AAAI 2020
66.92 79.62 61.53 86.86 45.36 71.45
4
Sep 27, 2019		 HGN (single model)
Microsoft Dynamics 365 AI Research
Fang et al., 2019
66.07 79.36 60.33 87.33 43.57 71.03
5
Jul 29, 2019		 TAP 2 (ensemble)
IBM Research AI & IISc
 66.64 79.82 57.21 86.69 41.21 70.65
6
Oct 1, 2019		 EPS + BERT(wwm) (single model)
Anonymous
65.79 79.05 58.50 86.26 42.47 70.48
7
Jul 29, 2019		 TAP 2 (single model)
IBM Research AI & IISc
 64.99 78.59 55.47 85.57 39.77 69.12
8
May 31, 2019		 EPS + BERT(large) (single model)
Anonymous
63.29 76.36 58.25 85.60 41.39 67.92
9
Aug 31, 2019		 SAE (single model)
JD AI Research
Tu, Huang et al., AAAI 2020
60.36 73.58 56.93 84.63 38.81 64.96
10
Jun 13, 2019		 P-BERT (single model)
Anonymous
61.18 74.16 51.38 82.76 35.42 63.79
11
Sep 16, 2019		 LQR-net 2 + BERT-Base (single model)
Anonymous
60.20 73.78 56.21 84.09 36.56 63.68
12
Apr 11, 2019		 EPS + BERT (single model)
Anonymous
60.13 73.31 52.55 83.20 35.40 63.41
13
May 16, 2019		 PIPE (single model)
Anonymous
59.77 72.77 52.53 82.82 35.54 62.92
14
Dec 1, 2019		 SEval (single model)
Anonymous
61.87 74.37 45.73 80.50 33.32 62.73
15
Jun 8, 2019		 TAP (single model)
58.63 71.48 46.84 82.98 32.03 61.90
16
Aug 14, 2019		 SAQA (single model)
Anonymous
55.07 70.22 57.62 84.19 35.94 61.72
17
Sep 2, 2019		 MKGN (single model)
Anonymous
57.09 70.69 54.26 83.54 35.59 61.69
18
Apr 19, 2019		 GRN + BERT (single model)
Anonymous
55.12 68.98 52.55 84.06 32.88 60.31
19
Jun 19, 2019		 LQR-net + BERT-Base (single model)
Anonymous
57.20 70.66 50.20 82.42 31.18 59.99
20
Apr 22, 2019		 DFGN (single model)
Shanghai Jiao Tong University & ByteDance AI Lab
(Xiao, Qu, Qiu et al. ACL19)		 56.31 69.69 51.50 81.62 33.62 59.82
21
Nov 21, 2018		 QFE (single model)
NTT Media Intelligence Laboratories
(Nishida et al., ACL'19)
53.86 68.06 57.75 84.49 34.63 59.61
22
Apr 17, 2019		 LQR-net (ensemble)
Anonymous
55.19 69.55 47.15 82.42 28.42 58.86
23
Mar 4, 2019		 GRN (single model)
Anonymous
52.92 66.71 52.37 84.11 31.77 58.47
24
Mar 1, 2019		 DFGN + BERT (single model)
Anonymous
55.17 68.49 49.85 81.06 31.87 58.23
25
Mar 4, 2019		 BERT Plus (single model)
CIS Lab
55.84 69.76 42.88 80.74 27.13 58.23
26
May 18, 2019		 KGNN (single model)
Tsinghua University
(Ye et al., 2019)
50.81 65.75 38.74 76.79 22.40 52.82
27
Oct 10, 2018		 Baseline Model (single model)
Carnegie Mellon University, Stanford University, & Universite de Montreal
(Yang, Qi, Zhang, et al. 2018)		 45.60 59.02 20.32 64.49 10.83 40.16
-
Feb 3, 2020		 Unsupervised Decomposition (single model)
Facebook AI Research, New York University & University College London
Unsupervised Question Decomposition for Question Answering
66.33 79.34 N/A N/A N/A N/A
-
Sep 24, 2019		 ChainEx (single model)
UT Austin
(Chen et al., 2019)
61.20 74.11 N/A N/A N/A N/A
-
Feb 27, 2019		 DecompRC (single model)
University of Washington
(Min et al., ACL'18)		 55.20 69.63 N/A N/A N/A N/A
-
Apr 2, 2019		 MatrixRC (single model)
Anonymous
47.07 60.75 N/A N/A N/A N/A
Leaderboard (Fullwiki Setting)
In the fullwiki setting, a question-answering system must find the answer to a question in the scope of the entire Wikipedia. Similar to in the distractor setting, systems are evaluated on the accuracy of their answers (Ans) and the quality of the supporting facts they use to justify them (Sup).
Model Code Ans Sup Joint
EM F1 EM F1 EM F1
1
Mar 6, 2020		 DR model large (single model)
Anonymous
62.01 75.32 49.88 77.77 35.44 62.95
2
Feb 11, 2020		 GAT-ALBERT (single model)
Microsoft Dynamics 365 AI Research
59.74 71.41 51.03 77.37 37.92 62.26
3
Nov 6, 2019		 Robustly Fine-tuned Graph-based Recurrent Retriever (single model)
Salesforce Research & University of Washington
(Asai et al., ICLR 2020)		 60.04 72.96 49.08 76.41 35.35 61.18
4
Dec 1, 2019		 HGN-large + SemanticRetrievalMRS IR (single model)
Anonymous
57.85 69.93 51.01 76.82 37.17 60.74
5
Oct 7, 2019		 HGN + SemanticRetrievalMRS IR (single model)
Microsoft Dynamics 365 AI Research
Fang et al., 2019
56.71 69.16 49.97 76.39 35.63 59.86
6
Feb 13, 2020		 DR model (single model)
Anonymous
58.82 71.68 41.55 72.54 29.34 56.82
7
Dec 8, 2019		 Quark + SemanticRetrievalMRS IR (single model)
Anonymous
55.50 67.51 45.64 72.95 32.89 56.23
8
Sep 20, 2019		 Graph-based Recurrent Retriever (single model)
Anonymous
56.04 68.87 44.14 73.03 29.18 55.31
9
Sep 28, 2019		 MIR+EPS+BERT (single model)
Anonymous
52.86 64.79 42.75 72.00 31.19 54.75
10
Feb 4, 2020		 Transformer-XH-final(BERT-base) (single model)
University of Maryland, Microsoft AI & Research
(Zhao et al. ICLR 2020)		 51.60 64.07 40.91 71.42 26.14 51.29
11
Sep 21, 2019		 Transformer-XH (single model)
Anonymous
48.95 60.75 41.66 70.01 27.13 49.57
12
May 15, 2019		 SemanticRetrievalMRS (single model)
UNC-NLP
(Nie et al., EMNLP'2019)		 45.32 57.34 38.67 70.83 25.14 47.60
13
Feb 21, 2020		 DrKIT (single model)
Carnegie Mellon University, Google Research
(Dhingra et al, ICLR 2020)		 42.13 51.72 37.05 59.84 24.69 42.88
14
Jul 31, 2019		 Entity-centric BERT Pipeline (single model)
Anonymous
41.82 53.09 26.26 57.29 17.01 39.18
15
May 21, 2019		 GoldEn Retriever (single model)
Stanford University
(Qi et al., EMNLP-IJCNLP 2019)		 37.92 48.58 30.69 64.24 18.04 39.13
16
Aug 14, 2019		 PR-Bert (single model)
KingSoft AI Lab
43.33 53.79 21.90 59.63 14.50 39.11
17
Dec 4, 2019		 SAFSr-Bert (single model)
Anonymous
39.35 51.40 24.21 58.54 13.34 37.00
18
Feb 21, 2019		 Cognitive Graph QA (single model)
Tsinghua KEG & Alibaba DAMO Academy
(Ding et al., ACL'19)		 37.12 48.87 22.82 57.69 12.42 34.92
19
Mar 5, 2019		 MUPPET (single model)
Technion
(Feldman and El-Yaniv, ACL'19)		 30.61 40.26 16.65 47.33 10.85 27.01
20
Apr 7, 2019		 GRN + BERT (single model)
Anonymous
29.87 39.14 13.16 49.67 8.26 25.84
21
May 20, 2019		 Entity-centric IR (single model)
Anonymous
35.36 46.26 0.06 43.16 0.02 25.47
22
May 19, 2019		 KGNN (single model)
Tsinghua University
(Ye et al., 2019)
27.65 37.19 12.65 47.19 7.03 24.66
23
Aug 16, 2019		 SAQA (single model)
Anonymous
28.44 38.62 14.69 47.17 8.62 24.49
24
Mar 4, 2019		 GRN (single model)
Anonymous
27.34 36.48 12.23 48.75 7.40 23.55
25
Nov 25, 2018		 QFE (single model)
NTT Media Intelligence Laboratories
(Nishida et al., ACL'19)
28.66 38.06 14.20 44.35 8.69 23.10
26
Nov 29, 2019		 SAFSr_model (single model)
Anonymous
28.91 39.14 8.03 40.55 4.06 20.90
27
Oct 12, 2018		 Baseline Model (single model)
Carnegie Mellon University, Stanford University, & Universite de Montreal
(Yang, Qi, Zhang, et al. 2018)		 23.95 32.89 3.86 37.71 1.85 16.15
-
May 19, 2019		 TPReasoner w/o BERT (single model)
Anonymous
36.04 47.43 N/A N/A N/A N/A
-
Feb 28, 2019		 DecompRC (single model)
University of Washington
(Min et al., ACL'18)		 30.00 40.65 N/A N/A N/A N/A
-
Mar 3, 2019		 MultiQA (single model)
Anonymous
30.73 40.23 N/A N/A N/A N/A