Edit detection and parsing for transcribed speech

Eugene Charniak; Mark Johnson

doi:10.3115/1073336.1073352

Edit detection and parsing for transcribed speech

Eugene Charniak, Mark Johnson

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding contribution › peer-review

127 Citations (Scopus)

13 Downloads (Pure)

Abstract

We present a simple architecture for parsing transcribed speech in which an edited-word detector first removes such words from the sentence string, and then a standard statistical parser trained on transcribed speech parses the remaining words. The edit detector achieves a misclassification rate on edited words of 2.2%. (The NULL-model, which marks everything as not edited, has an error rate of 5.9%.) To evaluate our parsing results we introduce a new evaluation metric, the purpose of which is to make evaluation of a parse tree relatively indifferent to the exact tree position of EDITED nodes. By this metric the parser achieves 85.3% precision and 86.5% recall.

Original language	English
Title of host publication	Proceedings of the 2nd Meeting of the North American Chapter of the Association for Computational Linguistics
Subtitle of host publication	NAACL 2001
Place of Publication	Stroudsburg, PA
Publisher	Association for Computational Linguistics (ACL)
Pages	118-126
Number of pages	9
DOIs	https://doi.org/10.3115/1073336.1073352
Publication status	Published - 2001
Externally published	Yes
Event	Meeting of the North American Chapter of the Association for Computational Linguistics (2nd : 2001) - Pittsburgh, United States Duration: 1 Jun 2001 → 7 Jun 2001

Conference

Conference	Meeting of the North American Chapter of the Association for Computational Linguistics (2nd : 2001)
Abbreviated title	NAACL '01
Country/Territory	United States
City	Pittsburgh
Period	1/06/01 → 7/06/01

Bibliographical note

Copyright the Publisher 2001. 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.

Access to Document

10.3115/1073336.1073352

Publisher version (open access)Licence: CC BY-NC-SA

https://www.aclweb.org/anthology/N01-1016/

Cite this

@inproceedings{5cc5a8682d2f45b89691b0f39273952d,

title = "Edit detection and parsing for transcribed speech",

abstract = "We present a simple architecture for parsing transcribed speech in which an edited-word detector first removes such words from the sentence string, and then a standard statistical parser trained on transcribed speech parses the remaining words. The edit detector achieves a misclassification rate on edited words of 2.2\%. (The NULL-model, which marks everything as not edited, has an error rate of 5.9\%.) To evaluate our parsing results we introduce a new evaluation metric, the purpose of which is to make evaluation of a parse tree relatively indifferent to the exact tree position of EDITED nodes. By this metric the parser achieves 85.3\% precision and 86.5\% recall. ",

author = "Eugene Charniak and Mark Johnson",

note = "Copyright the Publisher 2001. 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. ; Meeting of the North American Chapter of the Association for Computational Linguistics (2nd : 2001), NAACL '01 ; Conference date: 01-06-2001 Through 07-06-2001",

year = "2001",

doi = "10.3115/1073336.1073352",

language = "English",

pages = "118--126",

booktitle = "Proceedings of the 2nd Meeting of the North American Chapter of the Association for Computational Linguistics",

publisher = "Association for Computational Linguistics (ACL)",

}

Charniak, E & Johnson, M 2001, Edit detection and parsing for transcribed speech. in Proceedings of the 2nd Meeting of the North American Chapter of the Association for Computational Linguistics: NAACL 2001. Association for Computational Linguistics (ACL), Stroudsburg, PA, pp. 118-126, Meeting of the North American Chapter of the Association for Computational Linguistics (2nd : 2001), Pittsburgh, Pennsylvania, United States, 1/06/01. https://doi.org/10.3115/1073336.1073352

Edit detection and parsing for transcribed speech. / Charniak, Eugene; Johnson, Mark.
Proceedings of the 2nd Meeting of the North American Chapter of the Association for Computational Linguistics: NAACL 2001. Stroudsburg, PA: Association for Computational Linguistics (ACL), 2001. p. 118-126.

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding contribution › peer-review

TY - GEN

T1 - Edit detection and parsing for transcribed speech

AU - Charniak, Eugene

AU - Johnson, Mark

N1 - Copyright the Publisher 2001. 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 - 2001

Y1 - 2001

N2 - We present a simple architecture for parsing transcribed speech in which an edited-word detector first removes such words from the sentence string, and then a standard statistical parser trained on transcribed speech parses the remaining words. The edit detector achieves a misclassification rate on edited words of 2.2%. (The NULL-model, which marks everything as not edited, has an error rate of 5.9%.) To evaluate our parsing results we introduce a new evaluation metric, the purpose of which is to make evaluation of a parse tree relatively indifferent to the exact tree position of EDITED nodes. By this metric the parser achieves 85.3% precision and 86.5% recall.

AB - We present a simple architecture for parsing transcribed speech in which an edited-word detector first removes such words from the sentence string, and then a standard statistical parser trained on transcribed speech parses the remaining words. The edit detector achieves a misclassification rate on edited words of 2.2%. (The NULL-model, which marks everything as not edited, has an error rate of 5.9%.) To evaluate our parsing results we introduce a new evaluation metric, the purpose of which is to make evaluation of a parse tree relatively indifferent to the exact tree position of EDITED nodes. By this metric the parser achieves 85.3% precision and 86.5% recall.

UR - https://www.scopus.com/pages/publications/85120620835

U2 - 10.3115/1073336.1073352

DO - 10.3115/1073336.1073352

M3 - Conference proceeding contribution

SP - 118

EP - 126

BT - Proceedings of the 2nd Meeting of the North American Chapter of the Association for Computational Linguistics

PB - Association for Computational Linguistics (ACL)

CY - Stroudsburg, PA

T2 - Meeting of the North American Chapter of the Association for Computational Linguistics (2nd : 2001)

Y2 - 1 June 2001 through 7 June 2001

ER -

Edit detection and parsing for transcribed speech

Abstract

Conference

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this