Improving the robustness to recognition errors in speech input question answering

Hideki Tsutsui, Toshihiko Manabe, Mika Fukui, Tetsuya Sakai, Hiroko Fujii, Koji Urata

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

1 Citation (Scopus)

Abstract

In our previous work, we developed a prototype of a speech-input help system for home appliances such as digital cameras and microwave ovens. Given a factoid question, the system performs textual question answering using the manuals as the knowledge source. Whereas, given a HOW question, it retrieves and plays a demonstration video. However, our first prototype suffered from speech recognition errors, especially when the Japanese interrogative phrases in factoid questions were misrecognized. We therefore propose a method for solving this problem, which complements a speech query transcript with an interrogative phrase selected from a pre-determined list. The selection process first narrows down candidate phrases based on co-occurrences within the manual text, and then computes the similarity between each candidate and the query transcript in terms of pronunciation. Our method improves the Mean Reciprocal Rank of top three answers from 0.429 to 0.597 for factoid questions.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages297-312
Number of pages16
Volume4182 LNCS
Publication statusPublished - 2006
Externally publishedYes
Event3rd Asia Information Retrieval Symposium, AIRS 2006 - Singapore
Duration: 2006 Oct 162006 Oct 18

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4182 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other3rd Asia Information Retrieval Symposium, AIRS 2006
CitySingapore
Period06/10/1606/10/18

Fingerprint

Question Answering
Prototype
Query
Robustness
Microwave ovens
Domestic appliances
Digital Camera
Digital cameras
Speech Recognition
Speech recognition
Microwave
Demonstrations
Complement
Microwaves
Speech
Recognition (Psychology)
Knowledge
Similarity
Text

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

Tsutsui, H., Manabe, T., Fukui, M., Sakai, T., Fujii, H., & Urata, K. (2006). Improving the robustness to recognition errors in speech input question answering. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4182 LNCS, pp. 297-312). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4182 LNCS).

Improving the robustness to recognition errors in speech input question answering. / Tsutsui, Hideki; Manabe, Toshihiko; Fukui, Mika; Sakai, Tetsuya; Fujii, Hiroko; Urata, Koji.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4182 LNCS 2006. p. 297-312 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4182 LNCS).

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

Tsutsui, H, Manabe, T, Fukui, M, Sakai, T, Fujii, H & Urata, K 2006, Improving the robustness to recognition errors in speech input question answering. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 4182 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 4182 LNCS, pp. 297-312, 3rd Asia Information Retrieval Symposium, AIRS 2006, Singapore, 06/10/16.
Tsutsui H, Manabe T, Fukui M, Sakai T, Fujii H, Urata K. Improving the robustness to recognition errors in speech input question answering. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4182 LNCS. 2006. p. 297-312. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Tsutsui, Hideki ; Manabe, Toshihiko ; Fukui, Mika ; Sakai, Tetsuya ; Fujii, Hiroko ; Urata, Koji. / Improving the robustness to recognition errors in speech input question answering. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4182 LNCS 2006. pp. 297-312 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{3f62f8c8365b4de58c2ee7d6ad687bd2,
title = "Improving the robustness to recognition errors in speech input question answering",
abstract = "In our previous work, we developed a prototype of a speech-input help system for home appliances such as digital cameras and microwave ovens. Given a factoid question, the system performs textual question answering using the manuals as the knowledge source. Whereas, given a HOW question, it retrieves and plays a demonstration video. However, our first prototype suffered from speech recognition errors, especially when the Japanese interrogative phrases in factoid questions were misrecognized. We therefore propose a method for solving this problem, which complements a speech query transcript with an interrogative phrase selected from a pre-determined list. The selection process first narrows down candidate phrases based on co-occurrences within the manual text, and then computes the similarity between each candidate and the query transcript in terms of pronunciation. Our method improves the Mean Reciprocal Rank of top three answers from 0.429 to 0.597 for factoid questions.",
author = "Hideki Tsutsui and Toshihiko Manabe and Mika Fukui and Tetsuya Sakai and Hiroko Fujii and Koji Urata",
year = "2006",
language = "English",
isbn = "3540457801",
volume = "4182 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "297--312",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Improving the robustness to recognition errors in speech input question answering

AU - Tsutsui, Hideki

AU - Manabe, Toshihiko

AU - Fukui, Mika

AU - Sakai, Tetsuya

AU - Fujii, Hiroko

AU - Urata, Koji

PY - 2006

Y1 - 2006

N2 - In our previous work, we developed a prototype of a speech-input help system for home appliances such as digital cameras and microwave ovens. Given a factoid question, the system performs textual question answering using the manuals as the knowledge source. Whereas, given a HOW question, it retrieves and plays a demonstration video. However, our first prototype suffered from speech recognition errors, especially when the Japanese interrogative phrases in factoid questions were misrecognized. We therefore propose a method for solving this problem, which complements a speech query transcript with an interrogative phrase selected from a pre-determined list. The selection process first narrows down candidate phrases based on co-occurrences within the manual text, and then computes the similarity between each candidate and the query transcript in terms of pronunciation. Our method improves the Mean Reciprocal Rank of top three answers from 0.429 to 0.597 for factoid questions.

AB - In our previous work, we developed a prototype of a speech-input help system for home appliances such as digital cameras and microwave ovens. Given a factoid question, the system performs textual question answering using the manuals as the knowledge source. Whereas, given a HOW question, it retrieves and plays a demonstration video. However, our first prototype suffered from speech recognition errors, especially when the Japanese interrogative phrases in factoid questions were misrecognized. We therefore propose a method for solving this problem, which complements a speech query transcript with an interrogative phrase selected from a pre-determined list. The selection process first narrows down candidate phrases based on co-occurrences within the manual text, and then computes the similarity between each candidate and the query transcript in terms of pronunciation. Our method improves the Mean Reciprocal Rank of top three answers from 0.429 to 0.597 for factoid questions.

UR - http://www.scopus.com/inward/record.url?scp=33751374103&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33751374103&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:33751374103

SN - 3540457801

SN - 9783540457800

VL - 4182 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 297

EP - 312

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -