Dense passage retrieval for open-domain question answering

Vladimir Karpukhin; Barlas Oğuz; Sewon Min; Patrick Lewis; Ledell Wu; Sergey Edunov; Danqi Chen; Wen Tau Yih

Dense passage retrieval for open-domain question answering

Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, Wen Tau Yih

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

Open-domain question answering relies on efficient passage retrieval to select candidate contexts, where traditional sparse vector space models, such as TF-IDF or BM25, are the de facto method. In this work, we show that retrieval can be practically implemented using dense representations alone, where embeddings are learned from a small number of questions and passages by a simple dual-encoder framework. When evaluated on a wide range of open-domain QA datasets, our dense retriever outperforms a strong LuceneBM25 system greatly by 9%-19% absolute in terms of top-20 passage retrieval accuracy, and helps our end-to-end QA system establish new state-of-the-art on multiple open-domain QA benchmarks.

Original language	English (US)
Title of host publication	EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference
Publisher	Association for Computational Linguistics (ACL)
Pages	6769-6781
Number of pages	13
ISBN (Electronic)	9781952148606
State	Published - 2020
Event	2020 Conference on Empirical Methods in Natural Language Processing, EMNLP 2020 - Virtual, Online Duration: Nov 16 2020 → Nov 20 2020

Publication series

Name	EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference

Conference

Conference	2020 Conference on Empirical Methods in Natural Language Processing, EMNLP 2020
City	Virtual, Online
Period	11/16/20 → 11/20/20

All Science Journal Classification (ASJC) codes

Information Systems
Computer Science Applications
Computational Theory and Mathematics

Cite this

Karpukhin, V., Oğuz, B., Min, S., Lewis, P., Wu, L., Edunov, S., Chen, D., & Yih, W. T. (2020). Dense passage retrieval for open-domain question answering. In EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference (pp. 6769-6781). (EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference). Association for Computational Linguistics (ACL).

Karpukhin, Vladimir ; Oğuz, Barlas ; Min, Sewon et al. / Dense passage retrieval for open-domain question answering. EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference. Association for Computational Linguistics (ACL), 2020. pp. 6769-6781 (EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference).

@inproceedings{965c00256ba94afbb5cb88796804495a,

title = "Dense passage retrieval for open-domain question answering",

abstract = "Open-domain question answering relies on efficient passage retrieval to select candidate contexts, where traditional sparse vector space models, such as TF-IDF or BM25, are the de facto method. In this work, we show that retrieval can be practically implemented using dense representations alone, where embeddings are learned from a small number of questions and passages by a simple dual-encoder framework. When evaluated on a wide range of open-domain QA datasets, our dense retriever outperforms a strong LuceneBM25 system greatly by 9%-19% absolute in terms of top-20 passage retrieval accuracy, and helps our end-to-end QA system establish new state-of-the-art on multiple open-domain QA benchmarks.",

author = "Vladimir Karpukhin and Barlas Oğuz and Sewon Min and Patrick Lewis and Ledell Wu and Sergey Edunov and Danqi Chen and Yih, {Wen Tau}",

note = "Publisher Copyright: {\textcopyright} 2020 Association for Computational Linguistics; 2020 Conference on Empirical Methods in Natural Language Processing, EMNLP 2020 ; Conference date: 16-11-2020 Through 20-11-2020",

year = "2020",

language = "English (US)",

series = "EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference",

publisher = "Association for Computational Linguistics (ACL)",

pages = "6769--6781",

booktitle = "EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference",

}

Karpukhin, V, Oğuz, B, Min, S, Lewis, P, Wu, L, Edunov, S, Chen, D & Yih, WT 2020, Dense passage retrieval for open-domain question answering. in EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference. EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference, Association for Computational Linguistics (ACL), pp. 6769-6781, 2020 Conference on Empirical Methods in Natural Language Processing, EMNLP 2020, Virtual, Online, 11/16/20.

Dense passage retrieval for open-domain question answering. / Karpukhin, Vladimir; Oğuz, Barlas; Min, Sewon et al.
EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference. Association for Computational Linguistics (ACL), 2020. p. 6769-6781 (EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Dense passage retrieval for open-domain question answering

AU - Karpukhin, Vladimir

AU - Oğuz, Barlas

AU - Min, Sewon

AU - Lewis, Patrick

AU - Wu, Ledell

AU - Edunov, Sergey

AU - Chen, Danqi

AU - Yih, Wen Tau

PY - 2020

Y1 - 2020

N2 - Open-domain question answering relies on efficient passage retrieval to select candidate contexts, where traditional sparse vector space models, such as TF-IDF or BM25, are the de facto method. In this work, we show that retrieval can be practically implemented using dense representations alone, where embeddings are learned from a small number of questions and passages by a simple dual-encoder framework. When evaluated on a wide range of open-domain QA datasets, our dense retriever outperforms a strong LuceneBM25 system greatly by 9%-19% absolute in terms of top-20 passage retrieval accuracy, and helps our end-to-end QA system establish new state-of-the-art on multiple open-domain QA benchmarks.

AB - Open-domain question answering relies on efficient passage retrieval to select candidate contexts, where traditional sparse vector space models, such as TF-IDF or BM25, are the de facto method. In this work, we show that retrieval can be practically implemented using dense representations alone, where embeddings are learned from a small number of questions and passages by a simple dual-encoder framework. When evaluated on a wide range of open-domain QA datasets, our dense retriever outperforms a strong LuceneBM25 system greatly by 9%-19% absolute in terms of top-20 passage retrieval accuracy, and helps our end-to-end QA system establish new state-of-the-art on multiple open-domain QA benchmarks.

UR - http://www.scopus.com/inward/record.url?scp=85118435873&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85118435873&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:85118435873

T3 - EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference

SP - 6769

EP - 6781

BT - EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference

PB - Association for Computational Linguistics (ACL)

T2 - 2020 Conference on Empirical Methods in Natural Language Processing, EMNLP 2020

Y2 - 16 November 2020 through 20 November 2020

ER -

Karpukhin V, Oğuz B, Min S, Lewis P, Wu L, Edunov S et al. Dense passage retrieval for open-domain question answering. In EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference. Association for Computational Linguistics (ACL). 2020. p. 6769-6781. (EMNLP 2020 - 2020 Conference on Empirical Methods in Natural Language Processing, Proceedings of the Conference).

Dense passage retrieval for open-domain question answering

Abstract

Publication series

Conference

All Science Journal Classification (ASJC) codes

Other files and links

Fingerprint

Cite this