Improved combinatory categorial grammar induction with boundary words and bayesian inference

Yun Huang; Min Zhang; Chew Lim Tan

Improved combinatory categorial grammar induction with boundary words and bayesian inference

Yun Huang^*
, Min Zhang
, Chew Lim Tan

^*Corresponding author for this work

National University of Singapore
Agency for Science, Technology and Research, Singapore

Research output: Contribution to conference › Paper › peer-review

2 Scopus citations

Abstract

Combinatory Categorial Grammar (CCG) is an expressive grammar formalism which is able to capture long-range dependencies. However, building large and wide-coverage treebanks for CCG is expensive and time-consuming. In this paper, we focus on the problem of unsupervised CCG induction from plain texts. Based on the baseline model in (Bisk and Hockenmaier, 2012), we propose following two improvements: (1) we utilize boundary part-of-speech (POS) tags to capture lexical information; (2) we perform nonparametric Bayesian inference based on the Pitman-Yor process to learn compact grammars. Experiments on English Penn treebank demonstrate the effectiveness of our boundary model and Bayesian learning.

Original language	English
Pages	1257-1274
Number of pages	18
State	Published - 2012
Externally published	Yes
Event	24th International Conference on Computational Linguistics, COLING 2012 - Mumbai, India Duration: 8 Dec 2012 → 15 Dec 2012

Conference

Conference	24th International Conference on Computational Linguistics, COLING 2012
Country/Territory	India
City	Mumbai
Period	8/12/12 → 15/12/12

Keywords

Bayesian model
Boundary words
Combinatory categorial grammar
Grammar induction

Cite this

@conference{eb7a8f1cab264e09b48a3564704a9f19,

title = "Improved combinatory categorial grammar induction with boundary words and bayesian inference",

abstract = "Combinatory Categorial Grammar (CCG) is an expressive grammar formalism which is able to capture long-range dependencies. However, building large and wide-coverage treebanks for CCG is expensive and time-consuming. In this paper, we focus on the problem of unsupervised CCG induction from plain texts. Based on the baseline model in (Bisk and Hockenmaier, 2012), we propose following two improvements: (1) we utilize boundary part-of-speech (POS) tags to capture lexical information; (2) we perform nonparametric Bayesian inference based on the Pitman-Yor process to learn compact grammars. Experiments on English Penn treebank demonstrate the effectiveness of our boundary model and Bayesian learning.",

keywords = "Bayesian model, Boundary words, Combinatory categorial grammar, Grammar induction",

author = "Yun Huang and Min Zhang and Tan, \{Chew Lim\}",

year = "2012",

language = "英语",

pages = "1257--1274",

note = "24th International Conference on Computational Linguistics, COLING 2012 ; Conference date: 08-12-2012 Through 15-12-2012",

}

TY - CONF

T1 - Improved combinatory categorial grammar induction with boundary words and bayesian inference

AU - Huang, Yun

AU - Zhang, Min

AU - Tan, Chew Lim

PY - 2012

Y1 - 2012

N2 - Combinatory Categorial Grammar (CCG) is an expressive grammar formalism which is able to capture long-range dependencies. However, building large and wide-coverage treebanks for CCG is expensive and time-consuming. In this paper, we focus on the problem of unsupervised CCG induction from plain texts. Based on the baseline model in (Bisk and Hockenmaier, 2012), we propose following two improvements: (1) we utilize boundary part-of-speech (POS) tags to capture lexical information; (2) we perform nonparametric Bayesian inference based on the Pitman-Yor process to learn compact grammars. Experiments on English Penn treebank demonstrate the effectiveness of our boundary model and Bayesian learning.

AB - Combinatory Categorial Grammar (CCG) is an expressive grammar formalism which is able to capture long-range dependencies. However, building large and wide-coverage treebanks for CCG is expensive and time-consuming. In this paper, we focus on the problem of unsupervised CCG induction from plain texts. Based on the baseline model in (Bisk and Hockenmaier, 2012), we propose following two improvements: (1) we utilize boundary part-of-speech (POS) tags to capture lexical information; (2) we perform nonparametric Bayesian inference based on the Pitman-Yor process to learn compact grammars. Experiments on English Penn treebank demonstrate the effectiveness of our boundary model and Bayesian learning.

KW - Bayesian model

KW - Boundary words

KW - Combinatory categorial grammar

KW - Grammar induction

UR - https://www.scopus.com/pages/publications/84876784574

M3 - 论文

AN - SCOPUS:84876784574

SP - 1257

EP - 1274

T2 - 24th International Conference on Computational Linguistics, COLING 2012

Y2 - 8 December 2012 through 15 December 2012

ER -

Improved combinatory categorial grammar induction with boundary words and bayesian inference

Abstract

Conference

Keywords

Other files and links

Fingerprint

Cite this