Lexicalized stochastic modeling of constraint-based grammars using log-linear measures and EM training

Stefan Riezler; Detlef Prescher; Jonas Kuhn; Mark Johnson

doi:10.3115/1075218.1075279

Lexicalized stochastic modeling of constraint-based grammars using log-linear measures and EM training

Stefan Riezler, Detlef Prescher, Jonas Kuhn, Mark Johnson

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

34 Downloads (Pure)

Abstract

We present a new approach to stochastic modeling of constraint-based grammars that is based on loglinear models and uses EM for estimation from unannotated data. The techniques are applied to an LFG grammar for German. Evaluation on an exact match task yields 86% precision for an ambiguity rate of 5.4, and 90% precision on a subcat frame match for an ambiguity rate of 25. Experimental comparison to training from a parsebank shows a 10% gain from EM training. Also, a new class-based grammar lexicalization is presented, showing a 10% gain over unlexicalized models.

Original language	English
Title of host publication	Proceedings of the 38th Annual Meeting of the Association for Computational Linguistics (ACL'00), Hong Kong
Place of Publication	Stroudsburg, PA
Publisher	Association for Computational Linguistics (ACL)
Pages	480-487
Number of pages	8
DOIs	https://doi.org/10.3115/1075218.1075279
Publication status	Published - 2000
Externally published	Yes
Event	Annual Meeting of the Association for Computational Linguistics (38th : 2000) - Hong Kong Duration: 1 Oct 2000 → 8 Oct 2000

Conference

Conference	Annual Meeting of the Association for Computational Linguistics (38th : 2000)
City	Hong Kong
Period	1/10/00 → 8/10/00

Bibliographical note

Copyright the Publisher 2000. 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

10.3115/1075218.1075279

Publisher version (open access)Final published version, 232 KBLicence: CC BY-NC-SA

Cite this

Riezler, S., Prescher, D., Kuhn, J., & Johnson, M. (2000). Lexicalized stochastic modeling of constraint-based grammars using log-linear measures and EM training. In Proceedings of the 38th Annual Meeting of the Association for Computational Linguistics (ACL'00), Hong Kong (pp. 480-487). Association for Computational Linguistics (ACL). https://doi.org/10.3115/1075218.1075279

@inproceedings{004ef7f232704c27bedcf6761117b521,

title = "Lexicalized stochastic modeling of constraint-based grammars using log-linear measures and EM training",

abstract = "We present a new approach to stochastic modeling of constraint-based grammars that is based on loglinear models and uses EM for estimation from unannotated data. The techniques are applied to an LFG grammar for German. Evaluation on an exact match task yields 86% precision for an ambiguity rate of 5.4, and 90% precision on a subcat frame match for an ambiguity rate of 25. Experimental comparison to training from a parsebank shows a 10% gain from EM training. Also, a new class-based grammar lexicalization is presented, showing a 10% gain over unlexicalized models.",

author = "Stefan Riezler and Detlef Prescher and Jonas Kuhn and Mark Johnson",

note = "Copyright the Publisher 2000. 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. ; Annual Meeting of the Association for Computational Linguistics (38th : 2000) ; Conference date: 01-10-2000 Through 08-10-2000",

year = "2000",

doi = "10.3115/1075218.1075279",

language = "English",

pages = "480--487",

booktitle = "Proceedings of the 38th Annual Meeting of the Association for Computational Linguistics (ACL'00), Hong Kong",

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

}

Riezler, S, Prescher, D, Kuhn, J & Johnson, M 2000, Lexicalized stochastic modeling of constraint-based grammars using log-linear measures and EM training. in Proceedings of the 38th Annual Meeting of the Association for Computational Linguistics (ACL'00), Hong Kong. Association for Computational Linguistics (ACL), Stroudsburg, PA, pp. 480-487, Annual Meeting of the Association for Computational Linguistics (38th : 2000), Hong Kong, 1/10/00. https://doi.org/10.3115/1075218.1075279

Lexicalized stochastic modeling of constraint-based grammars using log-linear measures and EM training. / Riezler, Stefan; Prescher, Detlef; Kuhn, Jonas et al.
Proceedings of the 38th Annual Meeting of the Association for Computational Linguistics (ACL'00), Hong Kong. Stroudsburg, PA: Association for Computational Linguistics (ACL), 2000. p. 480-487.

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

TY - GEN

T1 - Lexicalized stochastic modeling of constraint-based grammars using log-linear measures and EM training

AU - Riezler, Stefan

AU - Prescher, Detlef

AU - Kuhn, Jonas

AU - Johnson, Mark

N1 - Copyright the Publisher 2000. 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 - 2000

Y1 - 2000

N2 - We present a new approach to stochastic modeling of constraint-based grammars that is based on loglinear models and uses EM for estimation from unannotated data. The techniques are applied to an LFG grammar for German. Evaluation on an exact match task yields 86% precision for an ambiguity rate of 5.4, and 90% precision on a subcat frame match for an ambiguity rate of 25. Experimental comparison to training from a parsebank shows a 10% gain from EM training. Also, a new class-based grammar lexicalization is presented, showing a 10% gain over unlexicalized models.

AB - We present a new approach to stochastic modeling of constraint-based grammars that is based on loglinear models and uses EM for estimation from unannotated data. The techniques are applied to an LFG grammar for German. Evaluation on an exact match task yields 86% precision for an ambiguity rate of 5.4, and 90% precision on a subcat frame match for an ambiguity rate of 25. Experimental comparison to training from a parsebank shows a 10% gain from EM training. Also, a new class-based grammar lexicalization is presented, showing a 10% gain over unlexicalized models.

UR - https://www.aclweb.org/anthology/P00-1061

U2 - 10.3115/1075218.1075279

DO - 10.3115/1075218.1075279

M3 - Conference proceeding contribution

SP - 480

EP - 487

BT - Proceedings of the 38th Annual Meeting of the Association for Computational Linguistics (ACL'00), Hong Kong

PB - Association for Computational Linguistics (ACL)

CY - Stroudsburg, PA

T2 - Annual Meeting of the Association for Computational Linguistics (38th : 2000)

Y2 - 1 October 2000 through 8 October 2000

ER -

Lexicalized stochastic modeling of constraint-based grammars using log-linear measures and EM training

Abstract

Conference

Bibliographical note

Access to Document

Other files and links

Cite this