<?xml version="1.0" standalone="yes"?>
<Paper uid="W03-1007">
  <Title>Maximum Entropy Models for FrameNet Classification</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Recent work in the development of FrameNet, a large database of semantically annotated sentences, has laid the foundation for statistical approaches to the task of automatic semantic classification.</Paragraph>
    <Paragraph position="1"> The FrameNet project seeks to annotate a large subset of the British National Corpus with semantic information. Annotations are based on Frame Semantics (Fillmore, 1976), in which frames are defined as schematic representations of situations involving various frame elements such as participants, props, and other conceptual roles.</Paragraph>
    <Paragraph position="2"> In each FrameNet sentence, a single target predicate is identified and all of its relevant frame elements are tagged with their semantic role (e.g., Agent, Judge), their syntactic phrase type (e.g., NP, PP), and their grammatical function (e.g., external argument, object argument). Figure 1 shows an example of an annotated sentence and its appropriate semantic frame.</Paragraph>
    <Paragraph position="3"> She clapped her hands in inspiration.</Paragraph>
    <Paragraph position="4">  core frame elements and a sentence annotated with element type, phrase type, and grammatical function. As of its first release in June 2002, FrameNet has made available 49,000 annotated sentences.</Paragraph>
    <Paragraph position="5"> The release contains 99,000 annotated frame elements for 1462 distinct lexical predicates (927 verbs, 339 nouns, and 175 adjectives).</Paragraph>
    <Paragraph position="6"> While considerable in scale, the FrameNet database does not yet approach the magnitude of resources available for other NLP tasks. Each target predicate, for example, has on average only 30 sentences tagged. This data sparsity makes the task of learning a semantic classifier formidable, and increases the importance of the modeling framework that is employed.</Paragraph>
  </Section>
class="xml-element"></Paper>