<?xml version="1.0" standalone="yes"?>
<Paper uid="P06-1111">
  <Title>Prototype-Driven Grammar Induction</Title>
  <Section position="10" start_page="887" end_page="887" type="concl">
    <SectionTitle>
8 Conclusion
</SectionTitle>
    <Paragraph position="0"> We have shown that distributional prototype features can allow one to specify a target labeling scheme in a compact and declarative way. These features give substantial error reduction in labeled F1 measure for English and Chinese grammar induction. They also achieve the best reported unlabeled F1 measure.8 Another positive property of this approach is that it tries to reconcile the success of distributional clustering approaches to grammar induction (Clark, 2001; Klein and Manning, 2002), with the CFG tree models in the supervised literature (Collins, 1999). Most importantly, this is the first work, to the authors' knowl8The next highest results being 77.1 and 46.7 for English and Chinese respectively from Klein and Manning (2004).</Paragraph>
    <Paragraph position="1"> edge, which has learned CFGs in an unsupervised or semi-supervised setting and can parse natural language language text with any reasonable accuracy. null Acknowledgments We would like to thank the anonymous reviewers for their comments. This work is supported by a Microsoft / CITRIS grant and by an equipment donation from Intel.</Paragraph>
  </Section>
class="xml-element"></Paper>