<?xml version="1.0" standalone="yes"?>
<Paper uid="H94-1080">
  <Title>A One Pass Decoder Design For Large Vocabulary Recognition</Title>
  <Section position="6" start_page="409" end_page="409" type="concl">
    <SectionTitle>
5. CONCLUSIONS
</SectionTitle>
    <Paragraph position="0"> One-pass decoding has many advantages over multi-pass decoding if it can be accomplished with a similar amount of computation. This paper has described a method of decoding continuous speech using context-dependent hidden Markov models and long span language models in a single pass. The decoder is relatively simple and efficient and should scale well with increasing size of both the vocabulary and the language models.</Paragraph>
  </Section>
class="xml-element"></Paper>