<?xml version="1.0" standalone="yes"?>
<Paper uid="N06-2006">
  <Title>Class Model Adaptation for Speech Summarisation</Title>
  <Section position="8" start_page="23" end_page="23" type="concl">
    <SectionTitle>
7 Conclusion
</SectionTitle>
    <Paragraph position="0"> In this paper we have investigated linguistic model adaptation using different sources of data for an automatic speech summarisation system. Class models have proved to be much more robust than word models for this process, and relative improvements ranging from 6.0% to 22.2% were obtained on a variety of evaluation metrics on summaries generated from automatic speech recogniser transcriptions.</Paragraph>
    <Paragraph position="1"> Acknowledgements: The authors would like to thank M. W&amp;quot;olfel for the recogniser transcriptions and C. Hori for her work on two stage summarisation and gathering the TED corpus data. This work is supported by the 21st Century COE Programme.</Paragraph>
  </Section>
class="xml-element"></Paper>