<?xml version="1.0" standalone="yes"?>
<Paper uid="P05-1045">
  <Title>Incorporating Non-local Information into Information Extraction Systems by Gibbs Sampling</Title>
  <Section position="9" start_page="369" end_page="369" type="concl">
    <SectionTitle>
9 Conclusions
</SectionTitle>
    <Paragraph position="0"> We have shown that a constraint model can be effectively combined with an existing sequence model in a factored architecture to successfully impose various sorts of long distance constraints. Our model generalizes naturally to other statistical models and other tasks. In particular, it could in the future be applied to statistical parsing. Statistical context free grammars provide another example of statistical models which are restricted to limiting local structure, and which could benefit from modeling non-local structure.</Paragraph>
  </Section>
class="xml-element"></Paper>