<?xml version="1.0" standalone="yes"?>
<Paper uid="W03-1110">
  <Title>Issues in Preand Post-translation Document Expansion: Untranslatable Cognates and Missegmented Words</Title>
  <Section position="10" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> These factors together explain both the significant improvement for post-translation document expansion that our experiments illustrate in contrast to the much weaker effects of pre-translation expansion, and also the difference observed between the experimental results reported here and prior work on preand post-translation query expansion that has emphasized European language pairs. We have identified a key role for post-translation expansion in CLIR language pairs where trivial cognate matching is not possible, but explicit translation or transliteration is required. We have also identified limitations on pre-translation expansion due to corresponding gaps in segmentation, translation, and transcription resources. We believe that these findings will extend to other CLIR language combinations with comparable characteristics, including many other Asian languages.</Paragraph>
  </Section>
class="xml-element"></Paper>