<?xml version="1.0" standalone="yes"?>
<Paper uid="C04-1047">
  <Title>Using a Mixture of N-Best Lists from Multiple MT Systems in Rank-Sum-Based Confidence Measure for MT Outputs</Title>
  <Section position="6" start_page="0" end_page="0" type="concl">
    <SectionTitle>
5 Conclusions
</SectionTitle>
    <Paragraph position="0"> This paper addressed the problem of eliminating unsatisfactory outputs from MT systems. It proposed a method that eliminates unsatisfactory outputs by using an alternative RSCM based on a mixture of N-best lists from multiple MT systems. The authors compared the proposed and existing RSCMs in the framework of an elimination system. When the number of MT outputs both in the N-best list for the existing RSCM and in the mixture of N-best lists for the proposed RSCM is almost the same number, i.e. ten, in most cases, the proposed RSCM proved to work better than the existing RSCM on two non-SMT systems and to work as well as the existing RSCM on an SMT system.</Paragraph>
    <Paragraph position="1"> In the future, the authors will conduct the following experiments: (1) investigating how the proposed RSCM works when the size of the M-best lists is increased, and (2) seeing how the proposed RSCM influences the performance of the selection system.</Paragraph>
  </Section>
class="xml-element"></Paper>