Can a suit of armor conduct electricity?: A new dataset for open book question answering

We present a new kind of question answering dataset, OpenBookQA, modeled after open book exams for assessing human understanding of a subject. The open book that comes with our questions is a set of 1326 elementary level science facts. Roughly 6000 questions probe an understanding of these facts and...

Full description

Saved in:
Bibliographic Details
Main Authors: Mihaylov, Todor (Author) , Clark, Peter (Author) , Khot, Tushar (Author) , Sabharwal, Ashish (Author)
Format: Chapter/Article Conference Paper
Language:English
Published: October/November 2018
In: EMNLP 2018
Year: 2018, Pages: 2381-2391
DOI:10.18653/v1/D18-1260
Online Access:Verlag, kostenfrei, Volltext: https://doi.org/10.18653/v1/D18-1260
Verlag, kostenfrei, Volltext: https://aclanthology.org/D18-1260
Get full text
Author Notes:Todor Mihaylov, Peter Clark, Tushar Khot, Ashish Sabharwal

MARC

LEADER 00000caa a2200000 c 4500
001 1895436907
003 DE-627
005 20241205151749.0
007 cr uuu---uuuuu
008 240715s2018 xx |||||o 00| ||eng c
024 7 |a 10.18653/v1/D18-1260  |2 doi 
035 |a (DE-627)1895436907 
035 |a (DE-599)KXP1895436907 
035 |a (OCoLC)1475302470 
040 |a DE-627  |b ger  |c DE-627  |e rda 
041 |a eng 
084 |a 28  |2 sdnb 
100 1 |a Mihaylov, Todor  |d 1989-  |e VerfasserIn  |0 (DE-588)1208448382  |0 (DE-627)1694700429  |4 aut 
245 1 0 |a Can a suit of armor conduct electricity?  |b A new dataset for open book question answering  |c Todor Mihaylov, Peter Clark, Tushar Khot, Ashish Sabharwal 
264 1 |c October/November 2018 
300 |a 11 
336 |a Text  |b txt  |2 rdacontent 
337 |a Computermedien  |b c  |2 rdamedia 
338 |a Online-Ressource  |b cr  |2 rdacarrier 
500 |a Gesehen am 15.07.2024 
520 |a We present a new kind of question answering dataset, OpenBookQA, modeled after open book exams for assessing human understanding of a subject. The open book that comes with our questions is a set of 1326 elementary level science facts. Roughly 6000 questions probe an understanding of these facts and their application to novel situations. This requires combining an open book fact (e.g., metals conduct electricity) with broad common knowledge (e.g., a suit of armor is made of metal) obtained from other sources. While existing QA datasets over documents or knowledge bases, being generally self-contained, focus on linguistic understanding, OpenBookQA probes a deeper understanding of both the topic—in the context of common knowledge—and the language it is expressed in. Human performance on OpenBookQA is close to 92%, but many state-of-the-art pre-trained QA methods perform surprisingly poorly, worse than several simple neural baselines we develop. Our oracle experiments designed to circumvent the knowledge retrieval bottleneck demonstrate the value of both the open book and additional facts. We leave it as a challenge to solve the retrieval problem in this multi-hop setting and to close the large gap to human performance. 
700 1 |a Clark, Peter  |e VerfasserIn  |4 aut 
700 1 |a Khot, Tushar  |e VerfasserIn  |4 aut 
700 1 |a Sabharwal, Ashish  |e VerfasserIn  |4 aut 
773 0 8 |i Enthalten in  |a Conference on Empirical Methods in Natural Language Processing (2018 : Brüssel)  |t EMNLP 2018  |d Stroudsburg, PA : Association for Computational Linguistics (ACL), 2018  |g (2018), Seite 2381-2391  |h 1 Online-Ressource (cxvi, 5051 Seiten, 344,93 MB)  |w (DE-627)1040746101  |z 9781948087841  |7 nnam 
773 1 8 |g year:2018  |g pages:2381-2391  |g extent:11  |a Can a suit of armor conduct electricity? A new dataset for open book question answering 
787 0 8 |i Forschungsdaten  |a Mihaylov, Todor, 1989 -   |t Knowledge-enhanced neural networks for machine reading comprehension [source code and additional material]"  |d Heidelberg : Universität, 2024  |h 1 Online-Ressource (4 Files)  |w (DE-627)1895400953 
856 4 0 |u https://doi.org/10.18653/v1/D18-1260  |x Verlag  |x Resolving-System  |z kostenfrei  |3 Volltext 
856 4 0 |u https://aclanthology.org/D18-1260  |x Verlag  |z kostenfrei  |3 Volltext 
951 |a AR 
992 |a 20240715 
993 |a ConferencePaper 
994 |a 2018 
998 |g 1208448382  |a Mihaylov, Todor  |m 1208448382:Mihaylov, Todor  |d 90000  |d 90500  |e 90000PM1208448382  |e 90500PM1208448382  |k 0/90000/  |k 1/90000/90500/  |p 1  |x j 
999 |a KXP-PPN1895436907  |e 4551137618 
BIB |a Y 
JSO |a {"person":[{"roleDisplay":"VerfasserIn","display":"Mihaylov, Todor","role":"aut","family":"Mihaylov","given":"Todor"},{"display":"Clark, Peter","roleDisplay":"VerfasserIn","role":"aut","family":"Clark","given":"Peter"},{"role":"aut","roleDisplay":"VerfasserIn","display":"Khot, Tushar","given":"Tushar","family":"Khot"},{"role":"aut","display":"Sabharwal, Ashish","roleDisplay":"VerfasserIn","given":"Ashish","family":"Sabharwal"}],"title":[{"subtitle":"A new dataset for open book question answering","title":"Can a suit of armor conduct electricity?","title_sort":"Can a suit of armor conduct electricity?"}],"language":["eng"],"recId":"1895436907","type":{"media":"Online-Ressource","bibl":"chapter"},"note":["Gesehen am 15.07.2024"],"name":{"displayForm":["Todor Mihaylov, Peter Clark, Tushar Khot, Ashish Sabharwal"]},"id":{"eki":["1895436907"],"doi":["10.18653/v1/D18-1260"]},"origin":[{"dateIssuedDisp":"October/November 2018","dateIssuedKey":"2018"}],"relHost":[{"corporate":[{"role":"aut","roleDisplay":"VerfasserIn","display":"Conference on Empirical Methods in Natural Language Processing (2018, Brüssel)"},{"display":"Association for Computational Linguistics","roleDisplay":"Herausgebendes Organ","role":"isb"}],"language":["eng"],"recId":"1040746101","note":["\"Editors: Ellen Riloff, David Chiang, Hockenmaier Julia, Tsujii Jun'ichi\" - Startseite der Ressource","Literaturangaben"],"disp":"Conference on Empirical Methods in Natural Language Processing (2018 : Brüssel)EMNLP 2018","type":{"bibl":"book","media":"Online-Ressource"},"part":{"year":"2018","pages":"2381-2391","text":"(2018), Seite 2381-2391","extent":"11"},"titleAlt":[{"title":"Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing"}],"person":[{"role":"edt","roleDisplay":"HerausgeberIn","display":"Riloff, Ellen","given":"Ellen","family":"Riloff"},{"family":"Chiang","given":"David","display":"Chiang, David","roleDisplay":"HerausgeberIn","role":"edt"},{"given":"Julia","family":"Hockenmaier","role":"edt","display":"Hockenmaier, Julia","roleDisplay":"HerausgeberIn"}],"title":[{"title_sort":"EMNLP 2018","title":"EMNLP 2018","subtitle":"Brussels, Belgium, Oct. 31-Nov. 4"}],"physDesc":[{"noteIll":"Illustrationen","extent":"1 Online-Ressource (cxvi, 5051 Seiten, 344,93 MB)"}],"id":{"isbn":["9781948087841"],"eki":["1040746101"]},"origin":[{"dateIssuedDisp":"[2018]","dateIssuedKey":"2018","publisher":"Association for Computational Linguistics (ACL)","publisherPlace":"Stroudsburg, PA"}]}],"physDesc":[{"extent":"11 S."}]} 
SRT |a MIHAYLOVTOCANASUITOF2018