<?xml version="1.0" standalone="yes"?>
<Paper uid="W99-0617">
  <Title>I POS Tags and Decision Trees for Language Modeling</Title>
  <Section position="10" start_page="135" end_page="136" type="ackno">
    <SectionTitle>
7 Acknowledgments
</SectionTitle>
    <Paragraph position="0"> We wish to thank James Allen, Geraldine Damnati, Chaojun Liu, Xintian Wu, and Yonghong Yan. This research work was partially supported by NSF under grant IRI9623665, by the Intel Research Council, and by CNET France T616com, while the author was visiting there.</Paragraph>
    <Paragraph position="1"> aWe compared a word-backoff model that does not exclude any trigrams or bigrams based on thresholds. Hence, the word-backoff approach can produce much smaller models. We still need to contrast our decision tree approach with smaller backoff models.</Paragraph>
  </Section>
class="xml-element"></Paper>