<?xml version="1.0" standalone="yes"?>
<Paper uid="P02-1038">
  <Title>Discriminative Training and Maximum Entropy Models for Statistical Machine Translation</Title>
  <Section position="8" start_page="0" end_page="0" type="concl">
    <SectionTitle>
7 Conclusions
</SectionTitle>
    <Paragraph position="0"> We have presented a framework for statistical MT for natural languages, which is more general than the widely used source-channel approach. It allows a baseline MT system to be extended easily by adding new feature functions. We have shown that a base-line statistical MT system can be significantly improved using this framework.</Paragraph>
    <Paragraph position="1"> There are two possible interpretations for a statistical MT system structured according to the source-channel approach, hence including a model for Pr(eI1) and a model for Pr(fJ1 jeI1). We can interpret it as an approximation to the Bayes decision rule in Eq. 2 or as an instance of a direct maximum entropy model with feature functions logPr(eI1) and logPr(fJ1 jeI1). As soon as we want to use model scaling factors, we can only do this in a theoretically justified way using the second interpretation. Yet, the main advantage comes from the large number of additional possibilities that we obtain by using the second interpretation.</Paragraph>
    <Paragraph position="2"> An important open problem of this approach is the handling of complex features in search. An interesting question is to come up with features that allow an efficient handling using conventional dynamic programming search algorithms.</Paragraph>
    <Paragraph position="3"> In addition, it might be promising to optimize the parameters directly with respect to the error rate of the MT system as is suggested in the field of pattern and speech recognition (Juang et al., 1995; Schl&amp;quot;uter and Ney, 2001).</Paragraph>
  </Section>
class="xml-element"></Paper>