<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-3255">
  <Title>Efficient Decoding for Statistical Machine Translation with a Fully Expanded WFST Model</Title>
  <Section position="7" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusions
</SectionTitle>
    <Paragraph position="0"> We proposed a method to compile statistical models to achieve efficient decoding in a machine translation system. In our method, each statistical sub-model is represented by a WFST, and all submodels are composed beforehand. To reduce the ambiguity of the composed WFST, the states are merged according to the statistics of hypotheses while decoding. As a result, we reduced decoding time to approximately a48a100a99 a103 a101 of dynamic composition of submodels, which corresponds to the conventional approach. null In this paper, we applied the state merging method to a fully-expanded WFST and showed the effectiveness of this approach. However, the state merging method itself is general and independent of the fully-expanded WFST. We can apply this method to each submodel of machine translation.</Paragraph>
    <Paragraph position="1"> More generally, we can apply it to all WFST-like models, including HMMs.</Paragraph>
  </Section>
class="xml-element"></Paper>