<?xml version="1.0" standalone="yes"?>
<Paper uid="C02-1058">
  <Title>Unsupervised Word Sense Disambiguation Using Bilingual Comparable Corpora</Title>
  <Section position="9" start_page="0" end_page="0" type="concl">
    <SectionTitle>
7 Conclusion
</SectionTitle>
    <Paragraph position="0"> A method for word sense disambiguation using a bilingual comparable corpus together with sense definitions by translations into another language was developed.</Paragraph>
    <Paragraph position="1"> In this method, knowledge for WSD, i.e., sense-vs.clue correlation, is acquired in an unsupervised fashion as follows. First, statistically significant pairs of related words are extracted from the corpus of each language. Then, aligning pairs of related words translingually, the correlation between the senses of a polysemous word and the clues, i.e., the words related to the polysemous word, is calculated. In order to overcome both the problem of ambiguity in the translingual alignment of pairs of related words and that of disparity of topical coverage between corpora of different languages, an iterative algorithm for calculating the correlation was developed.</Paragraph>
    <Paragraph position="2"> WSD for each instance of the polysemous word is done by selecting the sense that maximizes the score, i.e., the sum of the correlations between each sense and the clues appearing in the context of the instance. An experiment using corpora of English and Japanese newspaper articles showed that the performance of the new method is promising: the applicability and precision of sense selection were 88.5% and 77.7%, respectively, averaged over 60 test polysemous words.</Paragraph>
    <Paragraph position="3"> Acknowledgments: This research was sponsored in part by the Telecommunications Advancement Organization of Japan.</Paragraph>
  </Section>
class="xml-element"></Paper>