<?xml version="1.0" standalone="yes"?>
<Paper uid="P01-1037">
  <Title>The Role of Lexico-Semantic Feedback in Open-Domain Textual Question-Answering</Title>
  <Section position="8" start_page="0" end_page="0" type="concl">
    <SectionTitle>
7 Conclusion
</SectionTitle>
    <Paragraph position="0"> This paper has presented a Q&amp;/A system that employs several feedback mechanisms that provide lexical and semantic alternations to the question keywords. By relying on large, open-domain linguistic resources such as WordNet we enabled a more precise approach of searching and mining answers from large collections of texts. Evaluations indicate that when all three feedback loops are enabled we reached an enhancement of almost 76% for short answers and 91% for long answers, respectively, over the case when there are no feedback loops. In addition, a small increase is produced by relying on cached answers of similar questions. Our results so far indicate that the usage of feedback loops that produce alternations is significantly more efficient than multi-word indexing or annotations of large corpora with predicate-argument information.</Paragraph>
  </Section>
class="xml-element"></Paper>