Stability and accuracy in Incremental Speech Recognition

Ethan O. Selfridge, Iker Arizmendi, Peter Heeman, Jason D. Williams

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

21 Citations (Scopus)

Abstract

Conventional speech recognition approaches usually wait until the user has finished talking before returning a recognition hypothesis. This results in spoken dialogue systems that are unable to react while the user is still speaking. Incremental Speech Recognition (ISR), where partial phrase results are returned during user speech, has been used to create more reactive systems. However, ISR output is unstable and so prone to revision as more speech is decoded. This paper tackles the problem of stability in ISR. We first present a method that increases the stability and accuracy of ISR output, without adding delay. Given that some revisions are unavoidable, we next present a pair of methods for predicting the stability and accuracy of ISR results. Taken together, we believe these approaches give ISR more utility for real spoken dialogue systems.

Original languageEnglish (US)
Title of host publicationProceedings of the SIGDIAL 2011 Conference: 12th Annual Meeting of the Special Interest Group on Discourse and Dialogue
Pages110-119
Number of pages10
StatePublished - 2011
Event12th Annual Meeting of the Special Interest Group on Discourse and Dialogue, SIGDIAL 2011 - Portland, OR, United States
Duration: Jun 17 2011Jun 18 2011

Other

Other12th Annual Meeting of the Special Interest Group on Discourse and Dialogue, SIGDIAL 2011
CountryUnited States
CityPortland, OR
Period6/17/116/18/11

Fingerprint

Speech Recognition
Speech recognition
Spoken Dialogue Systems
Reactive Systems
Output
Unstable
Partial

ASJC Scopus subject areas

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

Cite this

Selfridge, E. O., Arizmendi, I., Heeman, P., & Williams, J. D. (2011). Stability and accuracy in Incremental Speech Recognition. In Proceedings of the SIGDIAL 2011 Conference: 12th Annual Meeting of the Special Interest Group on Discourse and Dialogue (pp. 110-119)

Stability and accuracy in Incremental Speech Recognition. / Selfridge, Ethan O.; Arizmendi, Iker; Heeman, Peter; Williams, Jason D.

Proceedings of the SIGDIAL 2011 Conference: 12th Annual Meeting of the Special Interest Group on Discourse and Dialogue. 2011. p. 110-119.

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

Selfridge, EO, Arizmendi, I, Heeman, P & Williams, JD 2011, Stability and accuracy in Incremental Speech Recognition. in Proceedings of the SIGDIAL 2011 Conference: 12th Annual Meeting of the Special Interest Group on Discourse and Dialogue. pp. 110-119, 12th Annual Meeting of the Special Interest Group on Discourse and Dialogue, SIGDIAL 2011, Portland, OR, United States, 6/17/11.
Selfridge EO, Arizmendi I, Heeman P, Williams JD. Stability and accuracy in Incremental Speech Recognition. In Proceedings of the SIGDIAL 2011 Conference: 12th Annual Meeting of the Special Interest Group on Discourse and Dialogue. 2011. p. 110-119
Selfridge, Ethan O. ; Arizmendi, Iker ; Heeman, Peter ; Williams, Jason D. / Stability and accuracy in Incremental Speech Recognition. Proceedings of the SIGDIAL 2011 Conference: 12th Annual Meeting of the Special Interest Group on Discourse and Dialogue. 2011. pp. 110-119
@inproceedings{e843980d716e49a2bbf59560656f5611,
title = "Stability and accuracy in Incremental Speech Recognition",
abstract = "Conventional speech recognition approaches usually wait until the user has finished talking before returning a recognition hypothesis. This results in spoken dialogue systems that are unable to react while the user is still speaking. Incremental Speech Recognition (ISR), where partial phrase results are returned during user speech, has been used to create more reactive systems. However, ISR output is unstable and so prone to revision as more speech is decoded. This paper tackles the problem of stability in ISR. We first present a method that increases the stability and accuracy of ISR output, without adding delay. Given that some revisions are unavoidable, we next present a pair of methods for predicting the stability and accuracy of ISR results. Taken together, we believe these approaches give ISR more utility for real spoken dialogue systems.",
author = "Selfridge, {Ethan O.} and Iker Arizmendi and Peter Heeman and Williams, {Jason D.}",
year = "2011",
language = "English (US)",
isbn = "9781937284107",
pages = "110--119",
booktitle = "Proceedings of the SIGDIAL 2011 Conference: 12th Annual Meeting of the Special Interest Group on Discourse and Dialogue",

}

TY - GEN

T1 - Stability and accuracy in Incremental Speech Recognition

AU - Selfridge, Ethan O.

AU - Arizmendi, Iker

AU - Heeman, Peter

AU - Williams, Jason D.

PY - 2011

Y1 - 2011

N2 - Conventional speech recognition approaches usually wait until the user has finished talking before returning a recognition hypothesis. This results in spoken dialogue systems that are unable to react while the user is still speaking. Incremental Speech Recognition (ISR), where partial phrase results are returned during user speech, has been used to create more reactive systems. However, ISR output is unstable and so prone to revision as more speech is decoded. This paper tackles the problem of stability in ISR. We first present a method that increases the stability and accuracy of ISR output, without adding delay. Given that some revisions are unavoidable, we next present a pair of methods for predicting the stability and accuracy of ISR results. Taken together, we believe these approaches give ISR more utility for real spoken dialogue systems.

AB - Conventional speech recognition approaches usually wait until the user has finished talking before returning a recognition hypothesis. This results in spoken dialogue systems that are unable to react while the user is still speaking. Incremental Speech Recognition (ISR), where partial phrase results are returned during user speech, has been used to create more reactive systems. However, ISR output is unstable and so prone to revision as more speech is decoded. This paper tackles the problem of stability in ISR. We first present a method that increases the stability and accuracy of ISR output, without adding delay. Given that some revisions are unavoidable, we next present a pair of methods for predicting the stability and accuracy of ISR results. Taken together, we believe these approaches give ISR more utility for real spoken dialogue systems.

UR - http://www.scopus.com/inward/record.url?scp=84863229187&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84863229187&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84863229187

SN - 9781937284107

SP - 110

EP - 119

BT - Proceedings of the SIGDIAL 2011 Conference: 12th Annual Meeting of the Special Interest Group on Discourse and Dialogue

ER -