<?xml version="1.0" standalone="yes"?>
<Paper uid="W95-0105">
  <Title>Disambiguating Noun Groupings with Respect to WordNet Senses</Title>
  <Section position="7" start_page="66" end_page="66" type="concl">
    <SectionTitle>
5 Conclusions and Future Work
</SectionTitle>
    <Paragraph position="0"> The results of the evaluation are exlremely encouraging, especially considering that disambiguating word senses to the level of fine-grainedness found in WordNet is quite a bit more difficult than disambiguation to the level of homographs (Hearst, 1991; Cowie et al., 1992). A note worth adding: it is not clear that the &amp;quot;exact match&amp;quot; criterion -- that is, evaluating algorithms by the percentage of exact matches of sense selection against a human-judged baseline -- is the right task. In particular, in many tasks it is at least as important to avoid inappropriate senses than to select exactly the right one. This would be the case in query expansion for information retrieval, for example, where indiscriminately adding inappropriate words to a query can degrade performance (Voorhees, 1994). The examples presented in Section 3 are encouraging in this regard: in addition to performing well at the task of assigning a high score to the best sense, it does a good job of assigning low scores to senses that are clearly inappropriate.</Paragraph>
    <Paragraph position="1"> Regardless of the criterion for success, the algorithm does need further evaluation. Immediate plans include a larger scale version of the experiment presented here, involving thesaurus classes, as well as a similarly designed evaluation of how the algorithm fares when presented with noun groups produced by distributional clustering. In addition, I plan to explore alternative measures of semantic similarity, for example an improved variant on simple path length that has been proposed by Leacock and Chodorow (1994).</Paragraph>
    <Paragraph position="2"> Ultimately, this algorithm is intended to be part of a suite of techniques used for disambiguating words in running text with respect to WordNet senses. I would argue that success at that task will require combining knowledge of the kind that WordNet provides, primarily about relatedness of meaning, with knowledge of the kind best provided by corpora, primarily about usage in context. The difficulty with the latter kind of knowledge is that, until now, the widespread success in characterizing lexical behavior in terms of distributional relationships has applied at the level of words -- indeed, word forms -- as opposed to senses.</Paragraph>
    <Paragraph position="3"> This paper represents a step toward getting as much leverage as possible out of work within that paradigm, and then using it to help determine relationships among word senses, which is really where the action is.</Paragraph>
  </Section>
class="xml-element"></Paper>