<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-0803">
  <Title>Parsing Word-Aligned Parallel Corpora in a Grammar Induction Context</Title>
  <Section position="7" start_page="23" end_page="23" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> We proposed a conceptually simple, yet efficient algorithm for synchronous parsing in a context where a word alignment can be assumed as given - for instance in a bootstrapping learning scenario. One of the two languages in synchronous parsing acts as the master language, providing the primary string span index, which is used as in classical Earley parsing.</Paragraph>
    <Paragraph position="1"> The second language contributes a bit vector as a secondary index, inspired by work on chart generation. Continuity assumptions make it possible to constrain the search space significantly, to the point that synchronous parsing for sentence pairs with few &amp;quot;NULL words&amp;quot; (which lack correspondents) may be faster than standard monolingual parsing. We discussed the complexity both theoretically and provided a quantitative evaluation based on a prototype implementation.</Paragraph>
    <Paragraph position="2"> The study we presented is part of the longer-term PTOLEMAIOS project. The next step is to apply the synchronous parsing algorithm with probabilistic synchronous grammars in grammar induction experiments on parallel corpora.</Paragraph>
  </Section>
class="xml-element"></Paper>