<?xml version="1.0" standalone="yes"?>
<Paper uid="W02-1033">
  <Title>An Analysis of the AskMSR Question-Answering System</Title>
  <Section position="7" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusions
</SectionTitle>
    <Paragraph position="0"> We have presented a novel approach to question-answering and carefully analyzed the contributions of each major system component, as well as analyzing what factors account for the majority of errors made by the AskMSR question answering system. In addition, we have demonstrated an approach to learning when the system is likely to answer a question incorrectly, allowing us to reach any desired rate of accuracy by not answering some portion of questions. We are currently exploring whether these techniques can be extended beyond short answer QA to more complex cases of information access.</Paragraph>
  </Section>
class="xml-element"></Paper>