<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-0854">
  <Title>KUNLP System in SENSEVAL-3</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> At SENSEVAL-3, we adopted an unsupervised approach based on WordNet and raw corpus, which does not require any sense tagged corpus. Word-Net specifies relationships among the meanings of words.</Paragraph>
    <Paragraph position="1"> Relatives of a word in WordNet are defined as words that have a relationship with it, e.g. they are synonyms, antonyms, superordinates (hypernyms), or subordinates (hyponyms). Relatives, especially those in a synonym class, usually have related meanings and tend to share similar contexts. Hence, some WordNet-based approaches extract relatives of each sense of a polysemous word from WordNet, collect example sentences of the relatives from a raw corpus, and learn the senses from the example sentences for WSD. Yarowsky (1992) first proposed this approach, but used International Roget's Thesaurus as a hierarchical lexical database instead of WordNet. However, the approach seems to suffer from examples irrelevant to the senses of a polysemous word since many of the relatives are polysemous. Leacock et al. (1998) attempted to exclude irrelevant or spurious examples by using only monosemous relatives in WordNet. However, some senses do not have short distance monosemous relatives through a relation such as synonym, child, and parent. A possible alternative of using only monosemous relatives in the long distance, however, is problematic because the longer the distance of two synsets in WordNet, the weaker the relationship between them. In other words, the monosemous relatives in the long distance may provide irrelevant examples for WSD.</Paragraph>
    <Paragraph position="2"> Our approach is somewhat similar to the Word-Net based approach of Leacock et al. (1998) in that it acquires relatives of a target word from WordNet and extracts co-occurrence frequencies of the relatives from a raw corpus, but our system uses polysemous as well as monosemous relatives. To avoid a negative effect of polysemous relatives on the co-occurrence frequency calculation, our system handles the example sentences of each relative separately instead of putting together the example sentences of all relatives into a pool. Also we devised our system to efficiently disambiguate senses of all words using only co-occurrence frequency between words.</Paragraph>
  </Section>
class="xml-element"></Paper>