<?xml version="1.0" standalone="yes"?>
<Paper uid="C02-2007">
  <Title>Building a Bilingual WordNet-Like Lexicon: the New Approach and Algorithms</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> As the processing of content information has nowadays become the center of NLP, a bilingual concept MRD is of increasingly great significance for IE, MT, WSD and the like. And it is for sure that the computational linguists would find such a lexicon indispensable and useful as semantic information when facing ambiguities in languages in their applications.</Paragraph>
    <Paragraph position="1"> At the same time, Princeton University's WordNet, after so many years' development, has exerted a profound influence on semantic lexicons [Vossen, 1998].</Paragraph>
    <Paragraph position="2"> When building a Chinese-English bilingual concept MRD, we must take the issue of compatibility with WordNet into account. In other words, for each English concept in WordNet, there should exist a corresponding Chinese concept in the bilingual lexicon and vice versa. Such a bilingual lexicon can offer better reusability and openness.</Paragraph>
    <Paragraph position="3"> The Institute of Computational Linguistics (ICL), Peking University, with this point of view, has launched the Project CCD (Chinese Concept Dictionary).</Paragraph>
    <Paragraph position="4"> The expectant CCD might be described as follows [Yu et al, 2001]: it should carry the main relations already defined in WordNet with more or less updates to reflect the reality of contemporary Chinese, and it should be a bilingual concept lexicon with the parallel Chinese-English concepts to be simultaneously included.</Paragraph>
    <Paragraph position="5"> Such a bilingual WordNet-like lexicon of Chinese-English concepts can largely meet our need of applications.</Paragraph>
    <Paragraph position="6"> However, it is by no means easy to build such a lexicon. It is quite obvious that there synchronously exist two ontologies in the same lexicon. One is in the English culture and the other is in the Chinese culture. As there might be different concepts and relations in each language, the mapping of the relevant concepts in different languages is inevitable. Also, the evolution of such a lexicon with passing of time, an issue linked closely to the mapping issue, is quite challenging.</Paragraph>
    <Paragraph position="7"> In conclusion, it's a quite demanding job to build such a lexicon, especially for the design of the approach and the realization of the developing tool. Any fruitful solution should give enough consideration to the complexity of these issues.</Paragraph>
  </Section>
class="xml-element"></Paper>