<?xml version="1.0" standalone="yes"?>
<Paper uid="P98-1066">
  <Title>A LAYERED APPROACH TO NLP-BASED INFORMATION RETRIEVAL</Title>
  <Section position="8" start_page="402" end_page="402" type="concl">
    <SectionTitle>
5 Future Directions
</SectionTitle>
    <Paragraph position="0"> Future work will concentrate on speed and space optimizations, and determining how subcomponents of this NLP capability can be incorporated into existing IR packages. This fine-grained NLP-based IR can also answer questions such as who, when, and where, so that the items retrieved can be more specifically targeted to user needs. The next step for caption-based systems will be to incorporate automatic disambiguation, so that captioners will not need to select a WordNet sense for each ambiguous word. In this auto-disambiguation investigation, it will be interesting to determine whether a specialized corpus, e.g. of photo captions, performs sense-tagging significantly better than a general-purpose corpus, such as the Brown corpus (Francis and Ku~era, 1979).</Paragraph>
  </Section>
class="xml-element"></Paper>