<?xml version="1.0" standalone="yes"?>
<Paper uid="P98-2248">
  <Title>Target Word Selection as Proximity in Semantic Space</Title>
  <Section position="4" start_page="0" end_page="1496" type="relat">
    <SectionTitle>
1.2 Related Work
</SectionTitle>
    <Paragraph position="0"> Dagan and Itai (1994) have also addressed the lexical selection problem from the TL point of view. Their algorithm uses information about local co-occurrence probabilities for all possible TL pairs of words that can result from translating each pair of words (verb/noun plus argument/modifier) in the SL sentence, and only  makes a decision if the preference is statistically significant. In work aimed at lexical choice in generation, Edmonds (1997) uses information about significant local co-occurrences to choose which of a set of synonyms is most typical in a given context. The present paper differs from these approaches in that local co-occurrence behaviour is not considered relevant, but rather an estimate of semantic relatedness between the TL context and each candidate translation.</Paragraph>
  </Section>
class="xml-element"></Paper>