<?xml version="1.0" standalone="yes"?>
<Paper uid="W04-0906">
  <Title>Question Answering Using Ontological Semantics</Title>
  <Section position="6" start_page="0" end_page="0" type="concl">
    <SectionTitle>
TMR browser
</SectionTitle>
    <Paragraph position="0"> We believe that integrating a comprehensive throughput system for an advanced application, even one in which some of the modules are still on a relatively small scale, is a very important kind of work in our field. It tackles real problems head on, same time, that does not imply a them). We are working g the knowledge (centered on the gical MEET-WITH script) to improve such In a separate article in preparation, we will go uch more detail about the reasoning process.</Paragraph>
    <Paragraph position="1"> obviously many additional issues, events. Our time resolution meaning procedures enable this; Assigning probabilities to inferences. For example, if two people were in the same room, the possibility of their meeting is much higher than if they were in the same country; Controlling the inference process.</Paragraph>
    <Paragraph position="2"> With regard to this last issue, the OntoSem ent provides a useful mechanism. In meant to constrain and without resorting to a rather defeatist - though quite common in today's NLP - claim that certain goals are infeasible.</Paragraph>
  </Section>
class="xml-element"></Paper>