<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-2934">
  <Title>Multi-lingual Dependency Parsing with Incremental Integer Linear Programming</Title>
  <Section position="7" start_page="229" end_page="229" type="concl">
    <SectionTitle>
5 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this work we presented a novel way of solving the linear model of McDonald et al. (2005a) for projective and non-projective parsing based on an incremental ILP approach. This allowed us to include additional linguistics constraints such as &amp;quot;a verb can only have one subject.&amp;quot; Due to time constraints we applied additional constraints to only four languages. For each one we gained better results than the baseline without constraints, however, this improvement was only marginal. This can be attributed to 4 main reasons: Firstly, the next best solution that fulfils the constraints was even worse (Chinese). Secondly, noisy POS tags caused coordination constraints to fail (Dutch). Thirdly, inference timed out (Chinese) and fourthly, constraints were not violated that often in the first place (Japanese).</Paragraph>
    <Paragraph position="1"> However, the effect of the first problem might be reduced by training with a higher k. The second problem could partly be overcome by using a better tagger or by a special treatment within the constraint handling for word types which are likely to be mistagged. The third problem could be avoidable by adding constraints during the branch and bound algorithm, avoiding the need to resolve the full problem &amp;quot;from scratch&amp;quot; for every constraint added. With these remedies significant improvements to the accuracy for some languages might be possible.</Paragraph>
  </Section>
class="xml-element"></Paper>