HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering. Yang, Z., Qi, P., Zhang, S., Bengio, Y., Cohen, W., Salakhutdinov, R., & Manning, C. D. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 2369–2380, Brussels, Belgium, October, 2018. Association for Computational Linguistics. ZSCC: 0000253
Paper doi abstract bibtex Existing question answering (QA) datasets fail to train QA systems to perform complex reasoning and provide explanations for answers. We introduce HotpotQA, a new dataset with 113k Wikipedia-based question-answer pairs with four key features: (1) the questions require finding and reasoning over multiple supporting documents to answer; (2) the questions are diverse and not constrained to any pre-existing knowledge bases or knowledge schemas; (3) we provide sentence-level supporting facts required for reasoning, allowing QA systems to reason with strong supervision and explain the predictions; (4) we offer a new type of factoid comparison questions to test QA systems' ability to extract relevant facts and perform necessary comparison. We show that HotpotQA is challenging for the latest QA systems, and the supporting facts enable models to improve performance and make explainable predictions.
@inproceedings{yang_hotpotqa_2018,
address = {Brussels, Belgium},
title = {{HotpotQA}: {A} {Dataset} for {Diverse}, {Explainable} {Multi}-hop {Question} {Answering}},
shorttitle = {{HotpotQA}},
url = {https://www.aclweb.org/anthology/D18-1259},
doi = {10.18653/v1/D18-1259},
abstract = {Existing question answering (QA) datasets fail to train QA systems to perform complex reasoning and provide explanations for answers. We introduce HotpotQA, a new dataset with 113k Wikipedia-based question-answer pairs with four key features: (1) the questions require finding and reasoning over multiple supporting documents to answer; (2) the questions are diverse and not constrained to any pre-existing knowledge bases or knowledge schemas; (3) we provide sentence-level supporting facts required for reasoning, allowing QA systems to reason with strong supervision and explain the predictions; (4) we offer a new type of factoid comparison questions to test QA systems' ability to extract relevant facts and perform necessary comparison. We show that HotpotQA is challenging for the latest QA systems, and the supporting facts enable models to improve performance and make explainable predictions.},
urldate = {2020-10-12},
booktitle = {Proceedings of the 2018 {Conference} on {Empirical} {Methods} in {Natural} {Language} {Processing}},
publisher = {Association for Computational Linguistics},
author = {Yang, Zhilin and Qi, Peng and Zhang, Saizheng and Bengio, Yoshua and Cohen, William and Salakhutdinov, Ruslan and Manning, Christopher D.},
month = oct,
year = {2018},
note = {ZSCC: 0000253},
pages = {2369--2380},
}
Downloads: 0
{"_id":"s6Si7cgNkqsAq43Eg","bibbaseid":"yang-qi-zhang-bengio-cohen-salakhutdinov-manning-hotpotqaadatasetfordiverseexplainablemultihopquestionanswering-2018","authorIDs":[],"author_short":["Yang, Z.","Qi, P.","Zhang, S.","Bengio, Y.","Cohen, W.","Salakhutdinov, R.","Manning, C. D."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","address":"Brussels, Belgium","title":"HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering","shorttitle":"HotpotQA","url":"https://www.aclweb.org/anthology/D18-1259","doi":"10.18653/v1/D18-1259","abstract":"Existing question answering (QA) datasets fail to train QA systems to perform complex reasoning and provide explanations for answers. We introduce HotpotQA, a new dataset with 113k Wikipedia-based question-answer pairs with four key features: (1) the questions require finding and reasoning over multiple supporting documents to answer; (2) the questions are diverse and not constrained to any pre-existing knowledge bases or knowledge schemas; (3) we provide sentence-level supporting facts required for reasoning, allowing QA systems to reason with strong supervision and explain the predictions; (4) we offer a new type of factoid comparison questions to test QA systems' ability to extract relevant facts and perform necessary comparison. We show that HotpotQA is challenging for the latest QA systems, and the supporting facts enable models to improve performance and make explainable predictions.","urldate":"2020-10-12","booktitle":"Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing","publisher":"Association for Computational Linguistics","author":[{"propositions":[],"lastnames":["Yang"],"firstnames":["Zhilin"],"suffixes":[]},{"propositions":[],"lastnames":["Qi"],"firstnames":["Peng"],"suffixes":[]},{"propositions":[],"lastnames":["Zhang"],"firstnames":["Saizheng"],"suffixes":[]},{"propositions":[],"lastnames":["Bengio"],"firstnames":["Yoshua"],"suffixes":[]},{"propositions":[],"lastnames":["Cohen"],"firstnames":["William"],"suffixes":[]},{"propositions":[],"lastnames":["Salakhutdinov"],"firstnames":["Ruslan"],"suffixes":[]},{"propositions":[],"lastnames":["Manning"],"firstnames":["Christopher","D."],"suffixes":[]}],"month":"October","year":"2018","note":"ZSCC: 0000253","pages":"2369–2380","bibtex":"@inproceedings{yang_hotpotqa_2018,\n\taddress = {Brussels, Belgium},\n\ttitle = {{HotpotQA}: {A} {Dataset} for {Diverse}, {Explainable} {Multi}-hop {Question} {Answering}},\n\tshorttitle = {{HotpotQA}},\n\turl = {https://www.aclweb.org/anthology/D18-1259},\n\tdoi = {10.18653/v1/D18-1259},\n\tabstract = {Existing question answering (QA) datasets fail to train QA systems to perform complex reasoning and provide explanations for answers. We introduce HotpotQA, a new dataset with 113k Wikipedia-based question-answer pairs with four key features: (1) the questions require finding and reasoning over multiple supporting documents to answer; (2) the questions are diverse and not constrained to any pre-existing knowledge bases or knowledge schemas; (3) we provide sentence-level supporting facts required for reasoning, allowing QA systems to reason with strong supervision and explain the predictions; (4) we offer a new type of factoid comparison questions to test QA systems' ability to extract relevant facts and perform necessary comparison. We show that HotpotQA is challenging for the latest QA systems, and the supporting facts enable models to improve performance and make explainable predictions.},\n\turldate = {2020-10-12},\n\tbooktitle = {Proceedings of the 2018 {Conference} on {Empirical} {Methods} in {Natural} {Language} {Processing}},\n\tpublisher = {Association for Computational Linguistics},\n\tauthor = {Yang, Zhilin and Qi, Peng and Zhang, Saizheng and Bengio, Yoshua and Cohen, William and Salakhutdinov, Ruslan and Manning, Christopher D.},\n\tmonth = oct,\n\tyear = {2018},\n\tnote = {ZSCC: 0000253},\n\tpages = {2369--2380},\n}\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n","author_short":["Yang, Z.","Qi, P.","Zhang, S.","Bengio, Y.","Cohen, W.","Salakhutdinov, R.","Manning, C. D."],"key":"yang_hotpotqa_2018","id":"yang_hotpotqa_2018","bibbaseid":"yang-qi-zhang-bengio-cohen-salakhutdinov-manning-hotpotqaadatasetfordiverseexplainablemultihopquestionanswering-2018","role":"author","urls":{"Paper":"https://www.aclweb.org/anthology/D18-1259"},"metadata":{"authorlinks":{}},"downloads":0,"html":""},"bibtype":"inproceedings","biburl":"https://bibbase.org/zotero/sarveshsoni","creationDate":"2020-01-27T02:13:33.974Z","downloads":0,"keywords":[],"search_terms":["hotpotqa","dataset","diverse","explainable","multi","hop","question","answering","yang","qi","zhang","bengio","cohen","salakhutdinov","manning"],"title":"HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering","year":2018,"dataSources":["hEoKh4ygEAWbAZ5iy","nxjWwW7fWbb5tfpKz","taWdMrienBzqHC2tC"]}