<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-0638">
  <Title>Exploiting Full Parsing Information to Label Semantic Roles Using an Ensemble of ME and SVM via Integer Linear Programming</Title>
  <Section position="6" start_page="235" end_page="235" type="concl">
    <SectionTitle>
4 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper, we add more full parsing features to argument classification models, and represent full parsing information as constraints in ILPs to resolve labeling inconsistencies. We also integrate two argument classification models, ME and SVM, by combining their argument classification results and applying them to the above-mentioned ILPs.</Paragraph>
    <Paragraph position="1"> The results show full parsing information increases the total F-score by 1.34%. The ensemble of SVM and ME also boosts the F-score by 0.77%. Finally, our system achieves an F-score of 76.53% in the development set and 76.38% in Test WSJ.</Paragraph>
  </Section>
class="xml-element"></Paper>