<?xml version="1.0" standalone="yes"?>
<Paper uid="W06-1617">
  <Title>Semantic Role Labeling of NomBank: A Maximum Entropy Approach</Title>
  <Section position="3" start_page="0" end_page="138" type="intro">
    <SectionTitle>
1 Introduction
Automatic Semantic Role Labeling (SRL) sys-
</SectionTitle>
    <Paragraph position="0"> tems, made possible by the availability of Prop-Bank (Kingsbury and Palmer, 2003; Palmer et al., 2005), and encouraged by evaluation efforts in (Carreras and Marquez, 2005; Litkowski, 2004), have been shown to accurately determine the argument structure of verb predicates.</Paragraph>
    <Paragraph position="1"> A successful PropBank-based SRL system would correctly determine that &amp;quot;Ben Bernanke&amp;quot; is the subject (labeled as ARG0 in PropBank) of predicate &amp;quot;replace&amp;quot;, and &amp;quot;Greenspan&amp;quot; is the object (labeled as ARG1): * Ben Bernanke replaced Greenspan as Fed chair.</Paragraph>
    <Paragraph position="2"> * Greenspan was replaced by Ben Bernanke as Fed chair.</Paragraph>
    <Paragraph position="3"> The recent release of NomBank (Meyers et al., 2004c; Meyers et al., 2004b), a databank that annotates argument structure for instances of common nouns in the Penn Treebank II corpus, made it possible to develop automatic SRL systems that analyze the argument structures of noun predicates. null Given the following two noun phrases and one sentence, a successful NomBank-based SRL system should label &amp;quot;Ben Bernanke&amp;quot; as the subject (ARG0) and &amp;quot;Greenspan&amp;quot; as the object (ARG1) of the noun predicate &amp;quot;replacement&amp;quot;.</Paragraph>
    <Paragraph position="4">  replacement.</Paragraph>
    <Paragraph position="5"> The ability to automatically analyze the argument structures of verb and noun predicates would greatly facilitate NLP tasks like question answering, information extraction, etc.</Paragraph>
    <Paragraph position="6"> This paper focuses on our efforts at building an accurate automatic NomBank-based SRL system. We study how techniques used in building PropBank SRL system can be transferred to developing NomBank SRL system. We also make NomBank-specific enhancements to our baseline system. Our implemented SRL system and experiments are based on the September 2005 release of NomBank (NomBank.0.8).</Paragraph>
    <Paragraph position="7"> The rest of this paper is organized as follows: Section 2 gives an overview of NomBank, Section 3 introduces the Maximum Entropy classification model, Section 4 introduces our features and feature selection strategy, Section 5 explains the experimental setup and presents the experimental results, Section 6 compares NomBank SRL to  beled in the style of NomBank PropBank SRL and discusses possible future research directions.</Paragraph>
  </Section>
class="xml-element"></Paper>