<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-0612">
  <Title>An Expectation Maximization Approach to Pronoun Resolution</Title>
  <Section position="9" start_page="94" end_page="94" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> We have demonstrated that unsupervised learning is possible for pronoun resolution. We achieve accuracy of 63% on an all-pronoun task, or 75% when a true antecedent is available to EM. There is now motivation to develop cleaner candidate lists and stronger probability models, with the hope of surpassing supervised techniques. For example, incorporating antecedent context, either at the sentence or document level, may boost performance. Furthermore, the lexicalized models learned in our system, especially the pronoun model, are potentially powerful features for any supervised pronoun resolution system.</Paragraph>
  </Section>
class="xml-element"></Paper>