<?xml version="1.0" standalone="yes"?>
<Paper uid="W03-1606">
  <Title>Normalization and Paraphrasing Using Symbolic Methods</Title>
  <Section position="8" start_page="0" end_page="0" type="concl">
    <SectionTitle>
5 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper, we have presented a methodology for extracting information using symbolic methods. Information extraction consists here in normalization of syntactic processing using both deep syntactic and morphological information as well as corpus specific knowledge. As the kind of corpus under consideration is very rich in reformulations, we were able to verify that our system could be used to detect paraphrases in the domain of the corpus. In fact, paraphrase detection can be seen as a side effect of normalization, as utterances conveying similar information are represented the same way. This is an on-going work but the first results we obtained for information extraction are really encouraging, although many improvements seem to be necessary. We foresee to continue our experiment applying our system on a different collection of texts from the same domain. We also plan to improve the current coverage of our system having in mind the results of the first evaluation.</Paragraph>
  </Section>
class="xml-element"></Paper>