<?xml version="1.0" standalone="yes"?>
<Paper uid="H94-1040">
  <Title>COMBINING KNOWLEDGE SOURCES TO REORDER N-BEST SPEECH HYPOTHESIS LISTS</Title>
  <Section position="7" start_page="219" end_page="220" type="concl">
    <SectionTitle>
4. CONCLUSIONS
</SectionTitle>
    <Paragraph position="0"> A simple and uniform architecture combines different knowledge sources to create an N-best preference method. The method can easily absorb new knowledge sources as they become available, and can be automatically trained. It is economical with regard to training material, since it makes use of both correct and incorrect recognizer hypotheses. It is in fact to be noted that over 80% of the discrimination  scores are negative, deriving from incorrect hypotheses. The apparent success of the method can perhaps most simply be explained by the fact that it attempts directly to model characteristic mistakes made by the recognizer. These are often idiosyncratic to a particular recognizer (or even to a particular version of a recognizer), and will not necessarily be easy to detect using more standard language models based on information derived from correct utterances only.</Paragraph>
    <Paragraph position="1"> We find the initial results described here encouraging, and in the next few months intend to extend them by training on larger amounts of data, refining existing knowledge sources, and adding new ones. In particular, we plan to investigate the possibility of improving the linguistic KSs by using partial linguistic analyses when a full analysis is not available. We are also experimenting with applying our methods to N-best lists that have first been rescored using normal class trigram models. Preliminary results indicate a proportional decrease of about 7% in the sentence error rate when syntactic variants of the reference sentence are counted as correct; this is significant according to the McNemar test. Only the finguistic KSs appear to contribute. We hope to be able to report these results in more detail at a later date.</Paragraph>
  </Section>
class="xml-element"></Paper>