<?xml version="1.0" standalone="yes"?>
<Paper uid="P04-1022">
  <Title>Collocation Translation Acquisition Using Monolingual Corpora</Title>
  <Section position="6" start_page="22" end_page="22" type="concl">
    <SectionTitle>
6 Conclusion and future work
</SectionTitle>
    <Paragraph position="0"> This paper proposes a novel method to train a triple translation model and extract collocation translations from two independent monolingual corpora. Evaluation results show that it outperforms the existing monolingual corpus based methods in triple translation, mainly due to the employment of EM algorithm in cross language translation probability estimation. By making use of the acquired triple translation model in two directions, promising results are achieved in collocation translation extraction.</Paragraph>
    <Paragraph position="1"> Our work also demonstrates the possibility of making full use of monolingual resources, such as corpora and parsers for bilingual tasks. This can help overcome the bottleneck of the lack of a large-scale bilingual corpus. This approach is also applicable to comparable corpora, which are also easier to access than bilingual corpora.</Paragraph>
    <Paragraph position="2"> In future work, we are interested in extending our method to solving the problem of non-compositional collocation translation. We are also interested in incorporating our triple translation model for sentence level translation.</Paragraph>
  </Section>
class="xml-element"></Paper>