<?xml version="1.0" standalone="yes"?>
<Paper uid="P05-3018">
  <Title>Word Alignment and Cross-Lingual Resource Acquisition [?]</Title>
  <Section position="6" start_page="71" end_page="71" type="concl">
    <SectionTitle>
4 Conclusion and Future Work
</SectionTitle>
    <Paragraph position="0"> In summary, we have presented an annotation environment for acquiring word alignments between English and Chinese as well as Part-Of-Speech tags for Chinese. The system is in place and the annotation process is underway.4 Once we have collected a medium-sized corpus, we will begin exploring different active learning techniques. Our goal is to find the best way to assign utility scores to the as-of-yet unlabeled sentences in order to obtain the greatest improvement in word alignment accuracy. Potential information useful for this task includes various measurements of the complexity of the sentence such as the rate of (automatic) alignments that are not one-to-one, the number of low-frequency words, and the number of potential language divergences (for example, many English verbs are nominalized in Chinese), and the co-occurrence of word pairs deemed to be unsure by the annotators in other contexts. Furthermore, we believe that the aggregate visualization tool will also help us uncover additional characteristics of potentially informative training examples.</Paragraph>
  </Section>
class="xml-element"></Paper>