Disfluency detection using auto-correlational neural networks

Research output: Chapter in Book/Report/Conference proceedingConference proceeding contributionResearchpeer-review

Abstract

In recent years, the natural language processing community has moved away from task-specific feature engineering, i.e., researchers discovering ad-hoc feature representations for various tasks, in favor of general-purpose methods that learn the input representation by themselves. However, state-of-the-art approaches to disfluency detection in spontaneous speech transcripts currently still depend on an array of hand-crafted features, and other representations derived from the output of pre-existing systems such as language models or dependency parsers. As an alternative, this paper proposes a simple yet effective model for automatic disfluency detection, called an auto-correlational neural network (ACNN). The model uses a convolutional neural network (CNN) and augments it with a new auto-correlation operator at the lowest layer that can capture the kinds of "rough copy" dependencies that are characteristic of repair disfluencies in speech. In experiments, the ACNN model outperforms the baseline CNN on a disfluency detection task with a 5% increase in f-score, which is close to the previous best result on this task.
LanguageEnglish
Title of host publicationProceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP 2018)
EditorsEllen Riloff, David Chiang, Julia Hockenmaier, Jun'ichi Tsujii
Place of PublicationStroudsburg
PublisherAssociation for Computational Linguistics (ACL)
Pages4610-4619
Number of pages10
ISBN (Electronic)9781948087841
Publication statusPublished - Nov 2018
Event2018 Conference on Empirical Methods in Natural Language Processing (EMNLP) - Brussels, Belgium
Duration: 31 Oct 20184 Nov 2018

Conference

Conference2018 Conference on Empirical Methods in Natural Language Processing (EMNLP)
CountryBelgium
CityBrussels
Period31/10/184/11/18

Fingerprint

Neural networks
Autocorrelation
Repair
Processing
Experiments

Bibliographical note

Copyright the Publisher 2018. Version archived for private and non-commercial use with the permission of the author/s and according to publisher conditions. For further rights please contact the publisher.

Cite this

Jamshid Lou, P., Anderson, P., & Johnson, M. (2018). Disfluency detection using auto-correlational neural networks. In E. Riloff, D. Chiang, J. Hockenmaier, & J. Tsujii (Eds.), Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP 2018) (pp. 4610-4619). [D18-1491] Stroudsburg: Association for Computational Linguistics (ACL).
Jamshid Lou, Paria ; Anderson, Peter ; Johnson, Mark. / Disfluency detection using auto-correlational neural networks. Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP 2018). editor / Ellen Riloff ; David Chiang ; Julia Hockenmaier ; Jun'ichi Tsujii. Stroudsburg : Association for Computational Linguistics (ACL), 2018. pp. 4610-4619
@inproceedings{1bfdb65166354df2abc3d7b1844a1ee4,
title = "Disfluency detection using auto-correlational neural networks",
abstract = "In recent years, the natural language processing community has moved away from task-specific feature engineering, i.e., researchers discovering ad-hoc feature representations for various tasks, in favor of general-purpose methods that learn the input representation by themselves. However, state-of-the-art approaches to disfluency detection in spontaneous speech transcripts currently still depend on an array of hand-crafted features, and other representations derived from the output of pre-existing systems such as language models or dependency parsers. As an alternative, this paper proposes a simple yet effective model for automatic disfluency detection, called an auto-correlational neural network (ACNN). The model uses a convolutional neural network (CNN) and augments it with a new auto-correlation operator at the lowest layer that can capture the kinds of {"}rough copy{"} dependencies that are characteristic of repair disfluencies in speech. In experiments, the ACNN model outperforms the baseline CNN on a disfluency detection task with a 5{\%} increase in f-score, which is close to the previous best result on this task.",
author = "{Jamshid Lou}, Paria and Peter Anderson and Mark Johnson",
note = "Copyright the Publisher 2018. Version archived for private and non-commercial use with the permission of the author/s and according to publisher conditions. For further rights please contact the publisher.",
year = "2018",
month = "11",
language = "English",
pages = "4610--4619",
editor = "Ellen Riloff and David Chiang and Julia Hockenmaier and Jun'ichi Tsujii",
booktitle = "Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP 2018)",
publisher = "Association for Computational Linguistics (ACL)",

}

Jamshid Lou, P, Anderson, P & Johnson, M 2018, Disfluency detection using auto-correlational neural networks. in E Riloff, D Chiang, J Hockenmaier & J Tsujii (eds), Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP 2018)., D18-1491, Association for Computational Linguistics (ACL), Stroudsburg, pp. 4610-4619, 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP), Brussels, Belgium, 31/10/18.

Disfluency detection using auto-correlational neural networks. / Jamshid Lou, Paria; Anderson, Peter; Johnson, Mark.

Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP 2018). ed. / Ellen Riloff; David Chiang; Julia Hockenmaier; Jun'ichi Tsujii. Stroudsburg : Association for Computational Linguistics (ACL), 2018. p. 4610-4619 D18-1491.

Research output: Chapter in Book/Report/Conference proceedingConference proceeding contributionResearchpeer-review

TY - GEN

T1 - Disfluency detection using auto-correlational neural networks

AU - Jamshid Lou, Paria

AU - Anderson, Peter

AU - Johnson, Mark

N1 - Copyright the Publisher 2018. Version archived for private and non-commercial use with the permission of the author/s and according to publisher conditions. For further rights please contact the publisher.

PY - 2018/11

Y1 - 2018/11

N2 - In recent years, the natural language processing community has moved away from task-specific feature engineering, i.e., researchers discovering ad-hoc feature representations for various tasks, in favor of general-purpose methods that learn the input representation by themselves. However, state-of-the-art approaches to disfluency detection in spontaneous speech transcripts currently still depend on an array of hand-crafted features, and other representations derived from the output of pre-existing systems such as language models or dependency parsers. As an alternative, this paper proposes a simple yet effective model for automatic disfluency detection, called an auto-correlational neural network (ACNN). The model uses a convolutional neural network (CNN) and augments it with a new auto-correlation operator at the lowest layer that can capture the kinds of "rough copy" dependencies that are characteristic of repair disfluencies in speech. In experiments, the ACNN model outperforms the baseline CNN on a disfluency detection task with a 5% increase in f-score, which is close to the previous best result on this task.

AB - In recent years, the natural language processing community has moved away from task-specific feature engineering, i.e., researchers discovering ad-hoc feature representations for various tasks, in favor of general-purpose methods that learn the input representation by themselves. However, state-of-the-art approaches to disfluency detection in spontaneous speech transcripts currently still depend on an array of hand-crafted features, and other representations derived from the output of pre-existing systems such as language models or dependency parsers. As an alternative, this paper proposes a simple yet effective model for automatic disfluency detection, called an auto-correlational neural network (ACNN). The model uses a convolutional neural network (CNN) and augments it with a new auto-correlation operator at the lowest layer that can capture the kinds of "rough copy" dependencies that are characteristic of repair disfluencies in speech. In experiments, the ACNN model outperforms the baseline CNN on a disfluency detection task with a 5% increase in f-score, which is close to the previous best result on this task.

UR - http://purl.org/au-research/grants/arc/DP160102156

M3 - Conference proceeding contribution

SP - 4610

EP - 4619

BT - Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP 2018)

A2 - Riloff, Ellen

A2 - Chiang, David

A2 - Hockenmaier, Julia

A2 - Tsujii, Jun'ichi

PB - Association for Computational Linguistics (ACL)

CY - Stroudsburg

ER -

Jamshid Lou P, Anderson P, Johnson M. Disfluency detection using auto-correlational neural networks. In Riloff E, Chiang D, Hockenmaier J, Tsujii J, editors, Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing (EMNLP 2018). Stroudsburg: Association for Computational Linguistics (ACL). 2018. p. 4610-4619. D18-1491