Implementing Variable Vectors in a CCG Parser

This article addresses problems that arise from the use of category variables T\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{do...

Full description

Saved in:
Bibliographic Details
Published inLogical Aspects of Computational Linguistics. Celebrating 20 Years of LACL (1996–2016) pp. 52 - 67
Main Authors Bekki, Daisuke, Kawazoe, Ai
Format Book Chapter
LanguageEnglish
Published Berlin, Heidelberg Springer Berlin Heidelberg 01.01.2016
SeriesLecture Notes in Computer Science
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:This article addresses problems that arise from the use of category variables T\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\varvec{T}}$$\end{document} in combinatory categorial grammars (CCGs), in particular, that they require extension of semantic languages with variable vectors in a form such as λx¯.M\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\lambda \overline{x}.M$$\end{document} or Mx¯\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$M\overline{x}$$\end{document}. As a solution to such problems, we introduce a technique for implementing variable vectors within the context of lightblue, a Japanese CCG parser implemented within the Haskell programming language with a dependent type semantics (DTS) representation.
Bibliography:My sincere thanks to Koji Mineshima and Ribeka Tanaka for many helpful discussions. I also thank the developers of JUMAN++ and KNP, in particular, Sadao Kurohashi, Daisuke Kawahara, Hajime Morita and Yuta Hayashibe, for sharing information. I also thank the anonymous reviewers of LACL2016 for many insightful comments. This research is partially supported by JST, CREST.
ISBN:3662538253
9783662538258
ISSN:0302-9743
1611-3349
DOI:10.1007/978-3-662-53826-5_4