Multi-Level Matching Networks for Text Matching

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Text matching aims to establish the matching relationship between two texts. It is an important operation in some information retrieval related tasks such as question duplicate detection, question answer- ing, and dialog systems. Bidirectional long short term memory (BiLSTM) coupled with attention mechanism has achieved state-of- the-art performance in text matching. A major limitation of existing works is that only high level contextualized word representations are utilized to obtain word level matching results without consider- ing other levels of word representations, thus resulting in incorrect matching decisions for cases where two words with different mean- ings are very close in high level contextualized word representation space. Therefore, instead of making decisions utilizing single level word representations, a multi-level matching network (MMN) is proposed in this paper for text matching, which utilizes multiple lev- els of word representations to obtain multiple word level matching results for final text level matching decision. Experimental results on two widely used benchmarks, SNLI and Scaitail, show that the proposed MMN achieves the state-of-the-art performance.
LanguageEnglish
Title of host publicationSIGIR 2019 - Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval
Pages949-952
Number of pages4
ISBN (Electronic)9781450361729
DOIs
Publication statusPublished - 18 Jul 2019

Fingerprint

Information retrieval
Decision making
Long short-term memory

Keywords

  • Attention
  • Multi-level matching network
  • Text matching
  • text matching
  • attention
  • multi-level matching network

Cite this

Xu, C., Lin, Z., Wang, H., & Wu, S. (2019). Multi-Level Matching Networks for Text Matching. In SIGIR 2019 - Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval (pp. 949-952) https://doi.org/10.1145/3331184.3331276
Xu, Chunlin ; Lin, Zhiwei ; Wang, Hui ; Wu, Shengli. / Multi-Level Matching Networks for Text Matching. SIGIR 2019 - Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. 2019. pp. 949-952
@inproceedings{a2623faf310849f88981fa6850171781,
title = "Multi-Level Matching Networks for Text Matching",
abstract = "Text matching aims to establish the matching relationship between two texts. It is an important operation in some information retrieval related tasks such as question duplicate detection, question answer- ing, and dialog systems. Bidirectional long short term memory (BiLSTM) coupled with attention mechanism has achieved state-of- the-art performance in text matching. A major limitation of existing works is that only high level contextualized word representations are utilized to obtain word level matching results without consider- ing other levels of word representations, thus resulting in incorrect matching decisions for cases where two words with different mean- ings are very close in high level contextualized word representation space. Therefore, instead of making decisions utilizing single level word representations, a multi-level matching network (MMN) is proposed in this paper for text matching, which utilizes multiple lev- els of word representations to obtain multiple word level matching results for final text level matching decision. Experimental results on two widely used benchmarks, SNLI and Scaitail, show that the proposed MMN achieves the state-of-the-art performance.",
keywords = "Attention, Multi-level matching network, Text matching, text matching, attention, multi-level matching network",
author = "Chunlin Xu and Zhiwei Lin and Hui Wang and Shengli Wu",
note = "This has an ISBN",
year = "2019",
month = "7",
day = "18",
doi = "10.1145/3331184.3331276",
language = "English",
isbn = "978-1-4503-6172-9",
pages = "949--952",
booktitle = "SIGIR 2019 - Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval",

}

Xu, C, Lin, Z, Wang, H & Wu, S 2019, Multi-Level Matching Networks for Text Matching. in SIGIR 2019 - Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. pp. 949-952. https://doi.org/10.1145/3331184.3331276

Multi-Level Matching Networks for Text Matching. / Xu, Chunlin; Lin, Zhiwei; Wang, Hui; Wu, Shengli.

SIGIR 2019 - Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. 2019. p. 949-952.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Multi-Level Matching Networks for Text Matching

AU - Xu, Chunlin

AU - Lin, Zhiwei

AU - Wang, Hui

AU - Wu, Shengli

N1 - This has an ISBN

PY - 2019/7/18

Y1 - 2019/7/18

N2 - Text matching aims to establish the matching relationship between two texts. It is an important operation in some information retrieval related tasks such as question duplicate detection, question answer- ing, and dialog systems. Bidirectional long short term memory (BiLSTM) coupled with attention mechanism has achieved state-of- the-art performance in text matching. A major limitation of existing works is that only high level contextualized word representations are utilized to obtain word level matching results without consider- ing other levels of word representations, thus resulting in incorrect matching decisions for cases where two words with different mean- ings are very close in high level contextualized word representation space. Therefore, instead of making decisions utilizing single level word representations, a multi-level matching network (MMN) is proposed in this paper for text matching, which utilizes multiple lev- els of word representations to obtain multiple word level matching results for final text level matching decision. Experimental results on two widely used benchmarks, SNLI and Scaitail, show that the proposed MMN achieves the state-of-the-art performance.

AB - Text matching aims to establish the matching relationship between two texts. It is an important operation in some information retrieval related tasks such as question duplicate detection, question answer- ing, and dialog systems. Bidirectional long short term memory (BiLSTM) coupled with attention mechanism has achieved state-of- the-art performance in text matching. A major limitation of existing works is that only high level contextualized word representations are utilized to obtain word level matching results without consider- ing other levels of word representations, thus resulting in incorrect matching decisions for cases where two words with different mean- ings are very close in high level contextualized word representation space. Therefore, instead of making decisions utilizing single level word representations, a multi-level matching network (MMN) is proposed in this paper for text matching, which utilizes multiple lev- els of word representations to obtain multiple word level matching results for final text level matching decision. Experimental results on two widely used benchmarks, SNLI and Scaitail, show that the proposed MMN achieves the state-of-the-art performance.

KW - Attention

KW - Multi-level matching network

KW - Text matching

KW - text matching

KW - attention

KW - multi-level matching network

UR - https://sigir.org/sigir2019/program/accepted/

UR - http://www.scopus.com/inward/record.url?scp=85073785181&partnerID=8YFLogxK

U2 - 10.1145/3331184.3331276

DO - 10.1145/3331184.3331276

M3 - Conference contribution

SN - 978-1-4503-6172-9

SP - 949

EP - 952

BT - SIGIR 2019 - Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval

ER -

Xu C, Lin Z, Wang H, Wu S. Multi-Level Matching Networks for Text Matching. In SIGIR 2019 - Proceedings of the 42nd International ACM SIGIR Conference on Research and Development in Information Retrieval. 2019. p. 949-952 https://doi.org/10.1145/3331184.3331276