<?xml version="1.0" standalone="yes"?>
<Paper uid="P04-1005">
  <Title>A TAG-based noisy channel model of speech repairs</Title>
  <Section position="5" start_page="0" end_page="0" type="concl">
    <SectionTitle>
5 Conclusion and further work
</SectionTitle>
    <Paragraph position="0"> This paper has proposed a novel noisy channel model of speech repairs and has used it to identify reparandum words. One of the advantages of probabilistic models is that they can be integrated with other probabilistic models in a principled way, and it would be interesting to investigate how to integrate this kind of model of speech repairs with probabilistic speech recognizers. null There are other kinds of joint models of reparandum and repair that may produce a better reparandum detection system. We have experimented with versions of the models described above based on POS bi-tag dependencies rather than word bigram dependencies, but with results very close to those presented here.</Paragraph>
    <Paragraph position="1"> Still, more sophisticated models may yield better performance.</Paragraph>
    <Paragraph position="2"> It would also be interesting to combine this probabilistic model of speech repairs with the word classi er approach of Charniak and Johnson (2001). That approach may do so well because many speech repairs are very short, involving only one or two words Shriberg and Stolcke (1998), so the reparandum, interregnum and repair are all contained in the surrounding word window used as features by the classier. On the other hand, the probabilistic model of repairs explored here seems to be most successful in identifying long repairs in which the reparandum and repair are similar enough to be unlikely to have been generated independently.</Paragraph>
    <Paragraph position="3"> Since the two approaches seem to have di erent strengths, a combined model may outperform both of them.</Paragraph>
  </Section>
class="xml-element"></Paper>