Automatic Detection of Hyperarticulated Speech

Eugenio Ribeiro, Fernando Batista, Isabel Trancoso, Ricardo Ribeiro, David de Matos

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingAcademicpeer-review

Abstract

Hyperarticulation is a speech adaptation that consists of adopting a clearer form of speech in an attempt to improve recognition levels. However, it has the opposite effect when talking to ASR systems, as they are not trained with such kind of speech. We present approaches for automatic detection of hyperarticulation, which can be used to improve the performance of spoken dialog systems. We performed experiments on Let’s Go data, using multiple feature sets and two classification approaches. Many relevant features are speaker dependent. Thus, we used the first turn in each dialog as the reference for the speaker, since it is typically not hyperarticulated. Our best results were above 80 % accuracy, which represents an improvement of at least 11.6 % points over previously obtained results on similar data. We also assessed the classifiers’ performance in scenarios where hyperarticulation is rare, achieving around 98 % accuracy using different confidence thresholds.
Original languageEnglish
Title of host publicationAdvances in Speech and Language Technologies for Iberian Languages
Subtitle of host publicationThird International Conference, IberSPEECH 2016, Lisbon, Portugal, November 23-25, 2016, Proceedings
EditorsAlberto Abad, Alfonso Ortega, António Teixeira, Carmen García Mateo, Carlos D. Martínez Hinarejos, Fernando Perdigão, Fernando Batista, Nuno Mamede
PublisherSpringer Nature Switzerland AG
Pages182-191
ISBN (Electronic)978-3-319-49169-1
ISBN (Print)978-3-319-49168-4
DOIs
Publication statusPublished - Nov 2016
Externally publishedYes
EventInternational Conference on Advances in Speech and Language Technologies for Iberian Languages - Lisbon, Portugal
Duration: 23 Nov 201625 Nov 2016
https://link.springer.com/conference/iberspeech

Publication series

NameLecture Notes in Computer Science book series
Volumevolume 10077
NameLecture Notes in Artificial Intelligence book sub series
Volumevolume 10077

Conference

ConferenceInternational Conference on Advances in Speech and Language Technologies for Iberian Languages
Abbreviated titleIberSPEECH 2016
CountryPortugal
CityLisbon
Period23/11/1625/11/16
Internet address

Fingerprint

Classifiers
Experiments

Keywords

  • Hyperarticulation
  • Speech
  • Let’s Go

Cite this

Ribeiro, E., Batista, F., Trancoso, I., Ribeiro, R., & de Matos, D. (2016). Automatic Detection of Hyperarticulated Speech. In A. Abad, A. Ortega, A. Teixeira, C. García Mateo, C. D. Martínez Hinarejos, F. Perdigão, F. Batista, ... N. Mamede (Eds.), Advances in Speech and Language Technologies for Iberian Languages: Third International Conference, IberSPEECH 2016, Lisbon, Portugal, November 23-25, 2016, Proceedings (pp. 182-191). (Lecture Notes in Computer Science book series; Vol. volume 10077), (Lecture Notes in Artificial Intelligence book sub series; Vol. volume 10077). Springer Nature Switzerland AG. https://doi.org/10.1007/978-3-319-49169-1_18
Ribeiro, Eugenio ; Batista, Fernando ; Trancoso, Isabel ; Ribeiro, Ricardo ; de Matos, David. / Automatic Detection of Hyperarticulated Speech. Advances in Speech and Language Technologies for Iberian Languages: Third International Conference, IberSPEECH 2016, Lisbon, Portugal, November 23-25, 2016, Proceedings. editor / Alberto Abad ; Alfonso Ortega ; António Teixeira ; Carmen García Mateo ; Carlos D. Martínez Hinarejos ; Fernando Perdigão ; Fernando Batista ; Nuno Mamede. Springer Nature Switzerland AG, 2016. pp. 182-191 (Lecture Notes in Computer Science book series). (Lecture Notes in Artificial Intelligence book sub series).
@inproceedings{7ad021dcb2e54e04a3270ef55be99649,
title = "Automatic Detection of Hyperarticulated Speech",
abstract = "Hyperarticulation is a speech adaptation that consists of adopting a clearer form of speech in an attempt to improve recognition levels. However, it has the opposite effect when talking to ASR systems, as they are not trained with such kind of speech. We present approaches for automatic detection of hyperarticulation, which can be used to improve the performance of spoken dialog systems. We performed experiments on Let’s Go data, using multiple feature sets and two classification approaches. Many relevant features are speaker dependent. Thus, we used the first turn in each dialog as the reference for the speaker, since it is typically not hyperarticulated. Our best results were above 80 {\%} accuracy, which represents an improvement of at least 11.6 {\%} points over previously obtained results on similar data. We also assessed the classifiers’ performance in scenarios where hyperarticulation is rare, achieving around 98 {\%} accuracy using different confidence thresholds.",
keywords = "Hyperarticulation, Speech, Let’s Go",
author = "Eugenio Ribeiro and Fernando Batista and Isabel Trancoso and Ricardo Ribeiro and {de Matos}, David",
year = "2016",
month = "11",
doi = "10.1007/978-3-319-49169-1_18",
language = "English",
isbn = "978-3-319-49168-4",
series = "Lecture Notes in Computer Science book series",
publisher = "Springer Nature Switzerland AG",
pages = "182--191",
editor = "Abad, {Alberto } and Ortega, {Alfonso } and Teixeira, {Ant{\'o}nio } and {Garc{\'i}a Mateo}, {Carmen } and {D. Mart{\'i}nez Hinarejos}, {Carlos } and Perdig{\~a}o, {Fernando } and Batista, {Fernando } and Mamede, {Nuno }",
booktitle = "Advances in Speech and Language Technologies for Iberian Languages",
address = "Switzerland",

}

Ribeiro, E, Batista, F, Trancoso, I, Ribeiro, R & de Matos, D 2016, Automatic Detection of Hyperarticulated Speech. in A Abad, A Ortega, A Teixeira, C García Mateo, C D. Martínez Hinarejos, F Perdigão, F Batista & N Mamede (eds), Advances in Speech and Language Technologies for Iberian Languages: Third International Conference, IberSPEECH 2016, Lisbon, Portugal, November 23-25, 2016, Proceedings. Lecture Notes in Computer Science book series, vol. volume 10077, Lecture Notes in Artificial Intelligence book sub series, vol. volume 10077, Springer Nature Switzerland AG, pp. 182-191, International Conference on Advances in Speech and Language Technologies for Iberian Languages, Lisbon, Portugal, 23/11/16. https://doi.org/10.1007/978-3-319-49169-1_18

Automatic Detection of Hyperarticulated Speech. / Ribeiro, Eugenio; Batista, Fernando; Trancoso, Isabel; Ribeiro, Ricardo; de Matos, David.

Advances in Speech and Language Technologies for Iberian Languages: Third International Conference, IberSPEECH 2016, Lisbon, Portugal, November 23-25, 2016, Proceedings. ed. / Alberto Abad; Alfonso Ortega; António Teixeira; Carmen García Mateo; Carlos D. Martínez Hinarejos; Fernando Perdigão; Fernando Batista; Nuno Mamede. Springer Nature Switzerland AG, 2016. p. 182-191 (Lecture Notes in Computer Science book series; Vol. volume 10077), (Lecture Notes in Artificial Intelligence book sub series; Vol. volume 10077).

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingAcademicpeer-review

TY - GEN

T1 - Automatic Detection of Hyperarticulated Speech

AU - Ribeiro, Eugenio

AU - Batista, Fernando

AU - Trancoso, Isabel

AU - Ribeiro, Ricardo

AU - de Matos, David

PY - 2016/11

Y1 - 2016/11

N2 - Hyperarticulation is a speech adaptation that consists of adopting a clearer form of speech in an attempt to improve recognition levels. However, it has the opposite effect when talking to ASR systems, as they are not trained with such kind of speech. We present approaches for automatic detection of hyperarticulation, which can be used to improve the performance of spoken dialog systems. We performed experiments on Let’s Go data, using multiple feature sets and two classification approaches. Many relevant features are speaker dependent. Thus, we used the first turn in each dialog as the reference for the speaker, since it is typically not hyperarticulated. Our best results were above 80 % accuracy, which represents an improvement of at least 11.6 % points over previously obtained results on similar data. We also assessed the classifiers’ performance in scenarios where hyperarticulation is rare, achieving around 98 % accuracy using different confidence thresholds.

AB - Hyperarticulation is a speech adaptation that consists of adopting a clearer form of speech in an attempt to improve recognition levels. However, it has the opposite effect when talking to ASR systems, as they are not trained with such kind of speech. We present approaches for automatic detection of hyperarticulation, which can be used to improve the performance of spoken dialog systems. We performed experiments on Let’s Go data, using multiple feature sets and two classification approaches. Many relevant features are speaker dependent. Thus, we used the first turn in each dialog as the reference for the speaker, since it is typically not hyperarticulated. Our best results were above 80 % accuracy, which represents an improvement of at least 11.6 % points over previously obtained results on similar data. We also assessed the classifiers’ performance in scenarios where hyperarticulation is rare, achieving around 98 % accuracy using different confidence thresholds.

KW - Hyperarticulation

KW - Speech

KW - Let’s Go

U2 - 10.1007/978-3-319-49169-1_18

DO - 10.1007/978-3-319-49169-1_18

M3 - Conference article in proceeding

SN - 978-3-319-49168-4

T3 - Lecture Notes in Computer Science book series

SP - 182

EP - 191

BT - Advances in Speech and Language Technologies for Iberian Languages

A2 - Abad, Alberto

A2 - Ortega, Alfonso

A2 - Teixeira, António

A2 - García Mateo, Carmen

A2 - D. Martínez Hinarejos, Carlos

A2 - Perdigão, Fernando

A2 - Batista, Fernando

A2 - Mamede, Nuno

PB - Springer Nature Switzerland AG

ER -

Ribeiro E, Batista F, Trancoso I, Ribeiro R, de Matos D. Automatic Detection of Hyperarticulated Speech. In Abad A, Ortega A, Teixeira A, García Mateo C, D. Martínez Hinarejos C, Perdigão F, Batista F, Mamede N, editors, Advances in Speech and Language Technologies for Iberian Languages: Third International Conference, IberSPEECH 2016, Lisbon, Portugal, November 23-25, 2016, Proceedings. Springer Nature Switzerland AG. 2016. p. 182-191. (Lecture Notes in Computer Science book series). (Lecture Notes in Artificial Intelligence book sub series). https://doi.org/10.1007/978-3-319-49169-1_18