<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-0835">
  <Title>A Recursive Statistical Translation Model[?]</Title>
  <Section position="11" start_page="206" end_page="206" type="concl">
    <SectionTitle>
11 Conclusions
</SectionTitle>
    <Paragraph position="0"> A new translation model has been presented. This model produces translations in a recursive way: the input sentence is divided in two parts, each is translated using the same procedure recursively and the translations are concatenated. The model has been used for finding the templates in a large vocabulary translation task. This involved using several heuristics to improve training time, including a method for splitting the input before training the models. Finally, the influence of using a stochastic dictionary together with the templates as a means of smoothing has been explored.</Paragraph>
  </Section>
class="xml-element"></Paper>