<?xml version="1.0" standalone="yes"?>
<Paper uid="W03-0411">
  <Title>Preposition Semantic Classification via PENN TREEBANK and FRAMENET</Title>
  <Section position="3" start_page="0" end_page="1" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> English prepositions convey important relations in text. When used as verbal adjuncts, they are the principle means of conveying semantic roles for the supporting entities described by the predicate. Prepositions are highly ambiguous. A typical collegiate dictionary has dozens of senses for each of the common prepositions. These senses tend to be closely related, in contrast to the other parts of speech where there might be a variety of distinct senses.</Paragraph>
    <Paragraph position="1"> Given the recent advances in word-sense disambiguation, due in part to SENSEVAL (Edmonds and Cotton, 2001), it would seem natural to apply the same basic approach to handling the disambiguation of prepositions. Of course, it is difficult to disambiguate prepositions at the granularity present in collegiate dictionaries, as illustrated later. Nonetheless, in certain cases this is feasible.</Paragraph>
    <Paragraph position="2"> We provide results for disambiguating prepositions at two different levels of granularity. The coarse granularity is more typical of earlier work in computational linguistics, such as the role inventory proposed by Fillmore (1968), including high-level roles such as instrument and location. Recently, systems have incorporated fine-grained roles, often specific to particular domains. For example, in the Cyc KB there are close to 200 different types of semantic roles. These range from high-level roles (e.g., beneficiaries) through medium-level roles (e.g., exchanges) to highly specialized roles (e.g., catalyst).</Paragraph>
    <Paragraph position="3">  Preposition classification using two different semantic role inventories are investigated in this paper, taking advantage of large annotated corpora.</Paragraph>
    <Paragraph position="4"> After providing background to the work in Section 2, experiments over the semantic role annotations are discussed in Section 3. The results over TREEBANK (Marcus et al., 1994) are covered first. Treebank include about a dozen high-level roles similar to Fillmore's. Next, experiments using the finer-grained semantic role annotations in FRAMENET version 0.75 (Fillmore et al., 2001) are  Part of the Cyc KB is freely available at www.opencyc.org. presented. FrameNet includes over 140 roles, approaching but not quite as specialized as Cyc's inventory. Section 4 follows with a comparison to related work, emphasizing work in broad-coverage preposition disambiguation.</Paragraph>
  </Section>
class="xml-element"></Paper>