<?xml version="1.0" standalone="yes"?>
<Paper uid="P06-1136">
  <Title>Reranking Answers for Definitional QA Using Language Modeling</Title>
  <Section position="4" start_page="1081" end_page="1081" type="relat">
    <SectionTitle>
2 Related Work
</SectionTitle>
    <Paragraph position="0"> Web information has been widely used for answer reranking and validation. For factoid QA task, AskMSR (Brill et al., 2001) ranks the answers by counting the occurrences of candidate answers returned from a search engine. Similarly, DIOGENE (Magnini et al., 2002) applies search engines to validate candidate answers.</Paragraph>
    <Paragraph position="1"> For definitional QA task, Lin (2002) presented an approach in which web-based answer reranking is combined with dictionary-based (e.g., WordNet) reranking, which leads to a 25% increase in mean reciprocal rank (MRR). Xu et al.</Paragraph>
    <Paragraph position="2"> (2003) proposed a statistical ranking method based on centroid vector (i.e., vector of words and frequencies) learned from the online encyclopedia (i.e., Wikipedia2) and the web. Candi-</Paragraph>
  </Section>
class="xml-element"></Paper>