<?xml version="1.0" standalone="yes"?>
<Paper uid="I05-5007">
  <Title>Automated Generalization of Phrasal Paraphrases from the Web*</Title>
  <Section position="10" start_page="54" end_page="54" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper, a novel method for automated generalization of paraphrase examples is proposed. This method is not dependent on the traditional limited texts instead it is based on the richness of the Web. It uses the multiple semantic codes to generalize a paraphrase example combing a semantic dictionary (Cilin (EV)). The experimental results proved that this representation method is reasonable and the generalized templates have a good precision and coverage.</Paragraph>
    <Paragraph position="1"> But this is just the beginning of the paraphrase examples generalization. And we simplify the problem in some aspects, such as we limited the number of the slot word in a paraphrase example, and we stipulate only the same word can be slot word. Also, we find that our templates are weak for those words or structures that don't exist in dictionary. Some methods in information extraction about named entities generalization can be used for reference in the future. Moreover, how to combine the semantic code with other representation forms together is also an interesting work.</Paragraph>
  </Section>
class="xml-element"></Paper>