<?xml version="1.0" standalone="yes"?>
<Paper uid="H92-1066">
  <Title>RAPID MATCH TRAINING FOR LARGE VOCABULARIES</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1. INTRODUCTION
</SectionTitle>
    <Paragraph position="0"> In this paper, we report on a new algorithm for building rapid match (prefiltering) models for Dragon's continuous speech recognizer. The rapid match module is intended to supply the recognizer with a relatively short list of word candidates at every point where the recognizer hypothesizes a new word may begin. To accomplish this, the rapid match module performs a quick but very approximate calculation using a short interval of acoustic data- usually no more than 240 milliseconds of speech and passes on to the recognizer a list of word candidates which can then be analyzed in detail.</Paragraph>
    <Paragraph position="1"> When the rapid match module for Dragon's continuous speech recognizer was first presented nearly two years ago \[1\], we evaluated its performance on a test corpus of mammography reports involving a vocabulary of under 1,000 words. At that time, the performance of the module was more than adequate to meet the demands of this recognition task. But as we move to larger vocabularies, the demands on rapid match have become greater at the same time that its role in recognition has become more crucial: if we hope to approach anything like real-time recognition on a large-vocabulary task using moderately priced personal computers, the recognizer can entertain *This work was sponsored by the Defense Advanced Research Projects Agency and was monitored by the Space and Naval Warfare Systems Command under contract N00039-86-C-0307.</Paragraph>
    <Paragraph position="2"> word hypotheses for only a tiny fraction of its complete vocabulary. Thus, not only must prefiltering provide models for more words, but those models must be capable of making finer distinctions.</Paragraph>
    <Paragraph position="3"> Until now, we had been generating rapid match models based on a single artificially constructed token representing the &amp;quot;average&amp;quot; behavior of each word. But working from a single token made it impossible to adequately model potential variability, and extensive adaptation of the models was necessary both to estimate variances and to adjust model parameters to new speakers. In our new training procedure, we instead build word models directly from hidden Markov models for each speaker's vocabulary. As reported below, these new models have allowed us to significantly improve prefiltering performance. null After a brief review of the rapid match module in the next section, we go on to describe in detail our new procedure for building rapid match models. Results from preliminary testing of these models using the Wall Street Journal recognition task are reported in section 4. We close with a discussion of the future directions we hope to explore.</Paragraph>
  </Section>
class="xml-element"></Paper>