Multi-domain spoken dialogue system with extensibility and robustness against speech recognition errors

Kazunori Komatani, Naoyuki Kanda, Mikio Nakano, Kazuhiro Nakadai, Hiroshi Tsujino, Tetsuya Ogata, Hiroshi G. Okuno

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

29 Citations (Scopus)

Abstract

We developed a multi-domain spoken dialogue system that can handle user requests across multiple domains. Such systems need to satisfy two requirements: extensibility and robustness against speech recognition errors. Extensibility is required to allow for the modification and addition of domains independent of other domains. Robustness against speech recognition errors is required because such errors are inevitable in speech recognition. However, the systems should still behave appropriately, even when their inputs are erroneous. Our system was constructed on an extensible architecture and is equipped with a robust and extensible domain selection method. Domain selection was based on three choices: (I) the previous domain, (II) the domain in which the speech recognition result can be accepted with the highest recognition score, and (III) other domains. With the third choice we newly introduced, our system can prevent dialogues from continuously being stuck in an erroneous domain. Our experimental results, obtained with 10 subjects, showed that our method reduced the domain selection errors by 18.3%, compared to a conventional method.

Original languageEnglish
Title of host publicationCOLING/ACL 2006 - SIGdial06: 7th SIGdial Workshop on Discourse and Dialogue, Proceedings of the Workshop
Pages9-17
Number of pages9
Publication statusPublished - 2006
Externally publishedYes
Event7th SIGdial Workshop on Discourse and Dialogue, SIGdial06 - Sydney, NSW
Duration: 2006 Jul 152006 Jul 16

Other

Other7th SIGdial Workshop on Discourse and Dialogue, SIGdial06
CitySydney, NSW
Period06/7/1506/7/16

Fingerprint

Spoken Dialogue Systems
Speech Recognition
Speech recognition
Robustness

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Computer Vision and Pattern Recognition
  • Human-Computer Interaction
  • Modelling and Simulation

Cite this

Komatani, K., Kanda, N., Nakano, M., Nakadai, K., Tsujino, H., Ogata, T., & Okuno, H. G. (2006). Multi-domain spoken dialogue system with extensibility and robustness against speech recognition errors. In COLING/ACL 2006 - SIGdial06: 7th SIGdial Workshop on Discourse and Dialogue, Proceedings of the Workshop (pp. 9-17)

Multi-domain spoken dialogue system with extensibility and robustness against speech recognition errors. / Komatani, Kazunori; Kanda, Naoyuki; Nakano, Mikio; Nakadai, Kazuhiro; Tsujino, Hiroshi; Ogata, Tetsuya; Okuno, Hiroshi G.

COLING/ACL 2006 - SIGdial06: 7th SIGdial Workshop on Discourse and Dialogue, Proceedings of the Workshop. 2006. p. 9-17.

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

Komatani, K, Kanda, N, Nakano, M, Nakadai, K, Tsujino, H, Ogata, T & Okuno, HG 2006, Multi-domain spoken dialogue system with extensibility and robustness against speech recognition errors. in COLING/ACL 2006 - SIGdial06: 7th SIGdial Workshop on Discourse and Dialogue, Proceedings of the Workshop. pp. 9-17, 7th SIGdial Workshop on Discourse and Dialogue, SIGdial06, Sydney, NSW, 06/7/15.
Komatani K, Kanda N, Nakano M, Nakadai K, Tsujino H, Ogata T et al. Multi-domain spoken dialogue system with extensibility and robustness against speech recognition errors. In COLING/ACL 2006 - SIGdial06: 7th SIGdial Workshop on Discourse and Dialogue, Proceedings of the Workshop. 2006. p. 9-17
Komatani, Kazunori ; Kanda, Naoyuki ; Nakano, Mikio ; Nakadai, Kazuhiro ; Tsujino, Hiroshi ; Ogata, Tetsuya ; Okuno, Hiroshi G. / Multi-domain spoken dialogue system with extensibility and robustness against speech recognition errors. COLING/ACL 2006 - SIGdial06: 7th SIGdial Workshop on Discourse and Dialogue, Proceedings of the Workshop. 2006. pp. 9-17
@inproceedings{9032edd526ca4b7b91c52c3bf68f241f,
title = "Multi-domain spoken dialogue system with extensibility and robustness against speech recognition errors",
abstract = "We developed a multi-domain spoken dialogue system that can handle user requests across multiple domains. Such systems need to satisfy two requirements: extensibility and robustness against speech recognition errors. Extensibility is required to allow for the modification and addition of domains independent of other domains. Robustness against speech recognition errors is required because such errors are inevitable in speech recognition. However, the systems should still behave appropriately, even when their inputs are erroneous. Our system was constructed on an extensible architecture and is equipped with a robust and extensible domain selection method. Domain selection was based on three choices: (I) the previous domain, (II) the domain in which the speech recognition result can be accepted with the highest recognition score, and (III) other domains. With the third choice we newly introduced, our system can prevent dialogues from continuously being stuck in an erroneous domain. Our experimental results, obtained with 10 subjects, showed that our method reduced the domain selection errors by 18.3{\%}, compared to a conventional method.",
author = "Kazunori Komatani and Naoyuki Kanda and Mikio Nakano and Kazuhiro Nakadai and Hiroshi Tsujino and Tetsuya Ogata and Okuno, {Hiroshi G.}",
year = "2006",
language = "English",
isbn = "193243271X",
pages = "9--17",
booktitle = "COLING/ACL 2006 - SIGdial06: 7th SIGdial Workshop on Discourse and Dialogue, Proceedings of the Workshop",

}

TY - GEN

T1 - Multi-domain spoken dialogue system with extensibility and robustness against speech recognition errors

AU - Komatani, Kazunori

AU - Kanda, Naoyuki

AU - Nakano, Mikio

AU - Nakadai, Kazuhiro

AU - Tsujino, Hiroshi

AU - Ogata, Tetsuya

AU - Okuno, Hiroshi G.

PY - 2006

Y1 - 2006

N2 - We developed a multi-domain spoken dialogue system that can handle user requests across multiple domains. Such systems need to satisfy two requirements: extensibility and robustness against speech recognition errors. Extensibility is required to allow for the modification and addition of domains independent of other domains. Robustness against speech recognition errors is required because such errors are inevitable in speech recognition. However, the systems should still behave appropriately, even when their inputs are erroneous. Our system was constructed on an extensible architecture and is equipped with a robust and extensible domain selection method. Domain selection was based on three choices: (I) the previous domain, (II) the domain in which the speech recognition result can be accepted with the highest recognition score, and (III) other domains. With the third choice we newly introduced, our system can prevent dialogues from continuously being stuck in an erroneous domain. Our experimental results, obtained with 10 subjects, showed that our method reduced the domain selection errors by 18.3%, compared to a conventional method.

AB - We developed a multi-domain spoken dialogue system that can handle user requests across multiple domains. Such systems need to satisfy two requirements: extensibility and robustness against speech recognition errors. Extensibility is required to allow for the modification and addition of domains independent of other domains. Robustness against speech recognition errors is required because such errors are inevitable in speech recognition. However, the systems should still behave appropriately, even when their inputs are erroneous. Our system was constructed on an extensible architecture and is equipped with a robust and extensible domain selection method. Domain selection was based on three choices: (I) the previous domain, (II) the domain in which the speech recognition result can be accepted with the highest recognition score, and (III) other domains. With the third choice we newly introduced, our system can prevent dialogues from continuously being stuck in an erroneous domain. Our experimental results, obtained with 10 subjects, showed that our method reduced the domain selection errors by 18.3%, compared to a conventional method.

UR - http://www.scopus.com/inward/record.url?scp=84857778881&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84857778881&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84857778881

SN - 193243271X

SN - 9781932432718

SP - 9

EP - 17

BT - COLING/ACL 2006 - SIGdial06: 7th SIGdial Workshop on Discourse and Dialogue, Proceedings of the Workshop

ER -