<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-1403">
  <Title>Sydney, July 2006. c(c)2006 Association for Computational Linguistics CCG Chart Realization from Disjunctive Inputs</Title>
  <Section position="3" start_page="0" end_page="12" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> In recent years, the generate-and-select paradigm of natural language generation has attracted increasing attention, particularly for the task of surface realization. In this paradigm, symbolic methods are used to generate a space of possible phrasings, and statistical methods are used to select one or more outputs from this space. To specify the desired paraphrase space, one may either provide an input logical form that underspecifies certain realization choices, or include explicit disjunctions in the input LF (or both). Our experience suggests that disjunctive LFs are an important capability, especially as one seeks to make grammars reusable across applications, and to employ domain-specific, sentence-level paraphrases (Barzilay and Lee, 2003).</Paragraph>
    <Paragraph position="1"> Prominent examples of surface realizers in the generate-and-select paradigm include Nitrogen/Halogen (Langkilde, 2000; Langkilde-Geary, 2002) and Fergus (Bangalore and Rambow, 2000).</Paragraph>
    <Paragraph position="2"> More recently, generate-and-select realizers in the chart realization tradition (Kay, 1996) have appeared, including the OpenCCG (White, 2004) and LinGO (Carroll and Oepen, 2005) realizers.</Paragraph>
    <Paragraph position="3"> Chart realizers make it possible to use the same reversible grammar for both parsing and realization, and employ well-defined methods of semantic composition to construct semantic representations that can properly represent the scope of logical operators.</Paragraph>
    <Paragraph position="4"> In the chart realization tradition, previous work has not generally supported disjunctive logical forms, with (Shemtov, 1997) as the only published exception (to the author's knowledge). Arguably, part of the reason that disjunctive LFs have not yet been embraced more broadly by those working on chart realization is that Shemtov's solution, while ingenious, is dauntingly complex. Looking beyond chart realizers, both Nitrogen/Halogen and Fergus support some forms of disjunctive input; however, in comparison to Shemtov's inputs, theirs are less expressive, in that they do not allowdisjunctionsacrossdifferentlevelsoftheinput null structure.</Paragraph>
    <Paragraph position="5"> As an alternative to Shemtov's method, this paper presents a chart realization algorithm for generating paraphrases from disjunctive logical forms that is more straightforward to implement, together with an initial case study of the algorithm's efficiency. As discussed in Section 5, the algorithm makes use of packed representations similar to those initially proposed by Shemtov, generalizing the approach in a way that avoids the problems that led Shemtov to reject his preliminary method. The algorithm is couched in the framework of Steedman's (2000) Combinatory Categorial Grammar (CCG) and has been implemented as an extension to the OpenCCG surface realizer.</Paragraph>
    <Paragraph position="6"> Though the algorithm is well suited to CCG, it is expected to be applicable to other constraint-based grammatical frameworks as well.</Paragraph>
    <Paragraph position="7">  (b) Semantic dependency graph for The design (is|'s) based on Villeroy and Boch's Funny Day series.</Paragraph>
    <Paragraph position="8"> (c) Disjunctive semantic dependency graph covering (a)(b), i.e. The design (is|'s) based on (the Funny Day (collection|series) by Villeroy and Boch  |Villeroy and Boch's Funny Day (collection|series)).</Paragraph>
  </Section>
class="xml-element"></Paper>