<?xml version="1.0" standalone="yes"?>
<Paper uid="P01-1022">
  <Title>Practical Issues in Compiling Typed Unification Grammars for Speech Recognition</Title>
  <Section position="8" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusions
</SectionTitle>
    <Paragraph position="0"> We have presented descriptions of two algorithms for approximating typed unification grammars with context-free grammars, and evaluated their performance during speech recognition. Initial results show that high levels of ambiguity coorelate with poor recognition performance, and that size of the resuling language model does not appear to directly coorelate with recognition performance.</Paragraph>
    <Paragraph position="1"> We have developed new techniques for further reducing the size and amount of ambiguity in these context-free grammars, but have so far met with mixed results.</Paragraph>
  </Section>
class="xml-element"></Paper>