<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-2906">
  <Title>Resolving and Generating Definite Anaphora by Modeling Hypernymy using Unlabeled Corpora</Title>
  <Section position="4" start_page="37" end_page="37" type="intro">
    <SectionTitle>
2 Related work
</SectionTitle>
    <Paragraph position="0"> There is a rich tradition of work using lexical and semantic resources for anaphora and coreference resolution. Several researchers have used WordNet as a lexical and semantic resource for certain types of bridging anaphora (Poesio et al., 1997; Meyer and Dale, 2002). WordNet has also been used as an important feature in machine learning of coreference resolution using supervised training data (Soon et al., 2001; Ng and Cardie, 2002). However, several researchers have reported that knowledge incorporated via WordNet is still insufficient for definite anaphora resolution. And of course, WordNet is not available for all languages and is missing inclusion of large segments of the vocabulary even for covered languages. Hence researchers have investigated use of corpus-based approaches to build a Word-Net like resource automatically (Hearst, 1992; Cara1The test examples were selected as follows: First, all the sentences containing definite NP &amp;quot;The Y&amp;quot; were extracted from the corpus. Then, the sentences containing instances of anaphoric definite NPs were kept and other cases of definite expressions (like existential NPs &amp;quot;The White House&amp;quot;,&amp;quot;The weather&amp;quot;)werediscarded. Fromthisanaphoricsetofsentences, 177 sentence instances covering 13 distinct hypernyms were randomly selected as the test set and annotated for the correct antecedent by human judges.</Paragraph>
    <Paragraph position="1"> ballo, 1999; Berland and Charniak, 1999). Also, several researchers have applied it to resolving different types of bridging anaphora (Clark, 1975).</Paragraph>
    <Paragraph position="2"> Poesio et al. (2002) have proposed extracting lexical knowledge about part-of relations using Hearst-style patterns and applied it to the task of resolving bridging references. Poesio et al. (2004) have suggested using Google as a source of computing lexical distance between antecedent and definite NP for mereological bridging references (references referring to parts of an object already introduced). Markert et al.</Paragraph>
    <Paragraph position="3"> (2003) have applied relations extracted from lexicosyntacticpatternssuchas'XandotherYs'forOther- null Anaphora (referential NPs with modifiers other or another) and for bridging involving meronymy.</Paragraph>
    <Paragraph position="4"> There has generally been a lack of work in the existing literature for automatically building lexical resources for definite anaphora resolution involving hyponyms relations such as presented in Example (1). However, this issue was recently addressed by Markert and Nissim (2005) by extending their work on Other-Anaphora using lexico syntactic pattern 'X and other Y's to antecedent selection for definite NP coreference. However, our task is more challenging since the anaphoric definite NPs in our test set include only hypernym anaphors without including the much simpler cases of headword repetition and other instances of string matching. For direct evaluation, we also implemented their corpus-based approach and compared it with our models on identical test data.</Paragraph>
    <Paragraph position="5">  Wealsodescribeandevaluateamechanismforcombining the knowledge obtained from WordNet and the six corpus-based approaches investigated here.</Paragraph>
    <Paragraph position="6">  Theresultingmodelsareabletoovercometheweaknesses of a WordNet-only model and substantially outperforms any of the individual models.</Paragraph>
  </Section>
class="xml-element"></Paper>