The impact of language models and loss functions on repair disfluency detection

Simon Zwarts*, Mark Johnson

*Corresponding author for this work

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

22 Citations (Scopus)
378 Downloads (Pure)

Abstract

Unrehearsed spoken language often contains disfluencies. In order to correctly interpret a spoken utterance, any such disfluencies must be identified and removed or otherwise dealt with. Operating on transcripts of speech which contain disfluencies, we study the effect of language model and loss function on the performance of a linear reranker that rescores the 25-best output of a noisychannel model. We show that language models trained on large amounts of non-speech data improve performance more than a languagemodel trained on amoremodest amount of speech data, and that optimising f-score rather than log loss improves disfluency detection performance. Our approach uses a log-linear reranker, operating on the top n analyses of a noisy channel model. We use large language models, introduce new features into this reranker and examine different optimisation strategies. We obtain a disfluency detection f-scores of 0.838 which improves upon the current state-of-theart.

Original languageEnglish
Title of host publicationProceedings of the 49th Annual Meeting of the Association for Computational Linguistics, ACL-HLT 2011
Place of PublicationPortland, OR
PublisherAssociation for Computational Linguistics (ACL)
Pages703-711
Number of pages9
Volume1
ISBN (Print)9781932432879
Publication statusPublished - 2011
Event49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, ACL-HLT 2011 - Portland, OR, United States
Duration: 19 Jun 201124 Jun 2011

Other

Other49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies, ACL-HLT 2011
CountryUnited States
CityPortland, OR
Period19/06/1124/06/11

Bibliographical note

Copyright the Publisher 2011. 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.

Fingerprint Dive into the research topics of 'The impact of language models and loss functions on repair disfluency detection'. Together they form a unique fingerprint.

  • Cite this

    Zwarts, S., & Johnson, M. (2011). The impact of language models and loss functions on repair disfluency detection. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics, ACL-HLT 2011 (Vol. 1, pp. 703-711). Portland, OR: Association for Computational Linguistics (ACL).