PCFG learning by nonterminal partition search

Research output: Chapter in Book/Conference proceeding with ISSN or ISBNConference contribution with ISSN or ISBNResearchpeer-review

Abstract

pcfg Learning by Partition Search is a general grammatical inference method for constructing, adapting and optimising pcfgs. Given a training corpus of examples from a language, a canonical grammar for the training corpus, and a parsing task, Partition Search pcfg Learning constructs a grammar that maximises performance on the parsing task and minimises grammar size. This paper describes Partition Search in detail, also providing theoretical background and a characterisation of the family of inference methods it belongs to. The paper also reports an example application to the task of building grammars for noun phrase extraction, a task that is crucial in many applications involving natu- ral language processing. In the experiments, Partition Search improves parsing performance by up to 21.45% compared to a general baseline and by up to 3.48% compared to a task-specific baseline, while reducing grammar size by up to 17.25%.
Original languageEnglish
Title of host publicationGrammatical Inference: Algorithms and Applications: Proceedings of the 6th International Colloquium: ICGI 2002
EditorsP. Adriaans, H. Fernau, M. van Zaanen
Place of PublicationBerlin, Germany
Pages14-27
Number of pages14
Volume2484/2
ISBN (Electronic)1611-3349
DOIs
Publication statusPublished - 1 Jan 2002
EventGrammatical Inference: Algorithms and Applications: Proceedings of the 6th International Colloquium: ICGI 2002 - Amsterdam, The Netherlands, September 23-25, 2002
Duration: 1 Jan 2002 → …

Publication series

NameLecture notes in computer science

Conference

ConferenceGrammatical Inference: Algorithms and Applications: Proceedings of the 6th International Colloquium: ICGI 2002
Period1/01/02 → …

Fingerprint

Processing
Experiments

Keywords

  • Partition search

Cite this

Belz, A. (2002). PCFG learning by nonterminal partition search. In P. Adriaans, H. Fernau, & M. van Zaanen (Eds.), Grammatical Inference: Algorithms and Applications: Proceedings of the 6th International Colloquium: ICGI 2002 (Vol. 2484/2, pp. 14-27). (Lecture notes in computer science). Berlin, Germany. https://doi.org/10.1.1.12.387
Belz, Anja. / PCFG learning by nonterminal partition search. Grammatical Inference: Algorithms and Applications: Proceedings of the 6th International Colloquium: ICGI 2002. editor / P. Adriaans ; H. Fernau ; M. van Zaanen. Vol. 2484/2 Berlin, Germany, 2002. pp. 14-27 (Lecture notes in computer science).
@inproceedings{b83c41ef3b5d4425bff235f78fd7a41c,
title = "PCFG learning by nonterminal partition search",
abstract = "pcfg Learning by Partition Search is a general grammatical inference method for constructing, adapting and optimising pcfgs. Given a training corpus of examples from a language, a canonical grammar for the training corpus, and a parsing task, Partition Search pcfg Learning constructs a grammar that maximises performance on the parsing task and minimises grammar size. This paper describes Partition Search in detail, also providing theoretical background and a characterisation of the family of inference methods it belongs to. The paper also reports an example application to the task of building grammars for noun phrase extraction, a task that is crucial in many applications involving natu- ral language processing. In the experiments, Partition Search improves parsing performance by up to 21.45{\%} compared to a general baseline and by up to 3.48{\%} compared to a task-specific baseline, while reducing grammar size by up to 17.25{\%}.",
keywords = "Partition search",
author = "Anja Belz",
year = "2002",
month = "1",
day = "1",
doi = "10.1.1.12.387",
language = "English",
isbn = "0302-9743",
volume = "2484/2",
series = "Lecture notes in computer science",
pages = "14--27",
editor = "P. Adriaans and H. Fernau and {van Zaanen}, M.",
booktitle = "Grammatical Inference: Algorithms and Applications: Proceedings of the 6th International Colloquium: ICGI 2002",

}

Belz, A 2002, PCFG learning by nonterminal partition search. in P Adriaans, H Fernau & M van Zaanen (eds), Grammatical Inference: Algorithms and Applications: Proceedings of the 6th International Colloquium: ICGI 2002. vol. 2484/2, Lecture notes in computer science, Berlin, Germany, pp. 14-27, Grammatical Inference: Algorithms and Applications: Proceedings of the 6th International Colloquium: ICGI 2002, 1/01/02. https://doi.org/10.1.1.12.387

PCFG learning by nonterminal partition search. / Belz, Anja.

Grammatical Inference: Algorithms and Applications: Proceedings of the 6th International Colloquium: ICGI 2002. ed. / P. Adriaans; H. Fernau; M. van Zaanen. Vol. 2484/2 Berlin, Germany, 2002. p. 14-27 (Lecture notes in computer science).

Research output: Chapter in Book/Conference proceeding with ISSN or ISBNConference contribution with ISSN or ISBNResearchpeer-review

TY - GEN

T1 - PCFG learning by nonterminal partition search

AU - Belz, Anja

PY - 2002/1/1

Y1 - 2002/1/1

N2 - pcfg Learning by Partition Search is a general grammatical inference method for constructing, adapting and optimising pcfgs. Given a training corpus of examples from a language, a canonical grammar for the training corpus, and a parsing task, Partition Search pcfg Learning constructs a grammar that maximises performance on the parsing task and minimises grammar size. This paper describes Partition Search in detail, also providing theoretical background and a characterisation of the family of inference methods it belongs to. The paper also reports an example application to the task of building grammars for noun phrase extraction, a task that is crucial in many applications involving natu- ral language processing. In the experiments, Partition Search improves parsing performance by up to 21.45% compared to a general baseline and by up to 3.48% compared to a task-specific baseline, while reducing grammar size by up to 17.25%.

AB - pcfg Learning by Partition Search is a general grammatical inference method for constructing, adapting and optimising pcfgs. Given a training corpus of examples from a language, a canonical grammar for the training corpus, and a parsing task, Partition Search pcfg Learning constructs a grammar that maximises performance on the parsing task and minimises grammar size. This paper describes Partition Search in detail, also providing theoretical background and a characterisation of the family of inference methods it belongs to. The paper also reports an example application to the task of building grammars for noun phrase extraction, a task that is crucial in many applications involving natu- ral language processing. In the experiments, Partition Search improves parsing performance by up to 21.45% compared to a general baseline and by up to 3.48% compared to a task-specific baseline, while reducing grammar size by up to 17.25%.

KW - Partition search

U2 - 10.1.1.12.387

DO - 10.1.1.12.387

M3 - Conference contribution with ISSN or ISBN

SN - 0302-9743

VL - 2484/2

T3 - Lecture notes in computer science

SP - 14

EP - 27

BT - Grammatical Inference: Algorithms and Applications: Proceedings of the 6th International Colloquium: ICGI 2002

A2 - Adriaans, P.

A2 - Fernau, H.

A2 - van Zaanen, M.

CY - Berlin, Germany

ER -

Belz A. PCFG learning by nonterminal partition search. In Adriaans P, Fernau H, van Zaanen M, editors, Grammatical Inference: Algorithms and Applications: Proceedings of the 6th International Colloquium: ICGI 2002. Vol. 2484/2. Berlin, Germany. 2002. p. 14-27. (Lecture notes in computer science). https://doi.org/10.1.1.12.387