<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-0830">
  <Title>Deploying Part-of-Speech Patterns to Enhance Statistical Phrase-Based Machine Translation Resources</Title>
  <Section position="5" start_page="165" end_page="165" type="concl">
    <SectionTitle>
4 Conclusion
</SectionTitle>
    <Paragraph position="0"> Syntactic and structural language information contained in a bilingual parallel corpus has been extracted and used to refine the translation probability values of a translation phrase table, using simple heuristics. The usability of the said translation table in statistical phrase-based machine translation has been tested in the shared task of the second track of the ACL 2005 Workshop on Building and Using Parallel Corpora. Findings suggest that using part-of-speech information to alter translation probabilities has had no significant effect upon translation performance. Further investigation is required to reveal how our approach can be optimized in order to produce significant performance improvement.</Paragraph>
  </Section>
class="xml-element"></Paper>