<?xml version="1.0" standalone="yes"?>
<Paper uid="N06-2036">
  <Title>Word Domain Disambiguation via Word Sense Disambiguation</Title>
  <Section position="2" start_page="0" end_page="141" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Word subject domains have been ubiquitously used in dictionaries to help human readers pinpoint the specific sense of a word by specifying technical usage, e.g. see &amp;quot;subject field codes&amp;quot; in Procter (1978). In computational linguistics, word subject domains have been widely used to improve the performance of machine translation systems. For example, in a review of commonly used features in automated translation, Mowatt (1999) reports that most of the machine translation systems surveyed made use of word subject domains. Word subject domains have also been used in information systems. For example, Sanfilippo (1998) describes a summarization system where subject domains provide users with useful conceptual parameters to tailor summary requests to a user's interest.</Paragraph>
    <Paragraph position="1"> Successful usage of word domains in applications such as machine translation and summarization is strongly dependent on the ability to assign the appropriate subject domain to a word in its context. Such an assignment requires a process of Word Domain Disambiguation (WDD) because the same word can often be assigned different subject domains out of context (e.g. the word partner can potentially be related to FINANCE or MARRIAGE).</Paragraph>
    <Paragraph position="2"> Interestingly enough, word subject domains have been widely used to improve the performance of Word Sense Disambiguation (WSD) algorithms (Wilks and Stevenson 1998, Magnini et al. 2001; Gliozzo et al. 2004). However, comparatively little effort has been devoted so far to the word domain disambiguation itself. The most notable exceptions are the work of Magnini and Strapparava (2000) and Suarez &amp; Palomar (2002). Both studies propose algorithms specific to the WDD task and have focused on the disambiguation of noun domains.</Paragraph>
    <Paragraph position="3"> In this paper we explore an alternative approach where word domain disambiguation is achieved via word sense disambiguation. Moreover, we extend the treatment of WDD to verbs and adjectives. Initial results show that this approach yield very strong results, suggesting that WDD can be addressed in terms of word sense disambiguation with no need of special purpose algorithms.</Paragraph>
  </Section>
class="xml-element"></Paper>