<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-0630">
  <Title>Hierarchical Semantic Role Labeling</Title>
  <Section position="3" start_page="0" end_page="201" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> For accomplishing the CoNLL 2005 Shared Task on Semantic Role Labeling (Carreras and M`arquez, 2005), we capitalized on our experience on the semantic shallow parsing by extending our system, widely experimented on PropBank and FrameNet (Giuglea and Moschitti, 2004) data, with a two-step boundary detection and a hierarchical argument classification strategy.</Paragraph>
    <Paragraph position="1"> Currently, the system can work in both basic and enhanced configuration. Given the parse tree of an input sentence, the basic system applies (1) a boundary classifier to select the nodes associated with correct arguments and (2) a multi-class labeler to assign the role type. For such models, we used some of the linear (e.g. (Gildea and Jurasfky, 2002; Pradhan et al., 2005)) and structural (Moschitti, 2004) features developed in previous studies.</Paragraph>
    <Paragraph position="2"> In the enhanced configuration, the boundary annotation is subdivided in two steps: a first pass in which we label argument boundary and a second pass in which we apply a simple heuristic to eliminate the argument overlaps. We have also tried some strategies to learn such heuristics automatically. In order to do this we used a tree kernel to classify the subtrees associated with correct predicate argument structures (see (Moschitti et al., 2005)). The rationale behind such an attempt was to exploit the correlation among potential arguments.</Paragraph>
    <Paragraph position="3"> Also, the role labeler is divided into two steps: (1) we assign to the arguments one out of four possible class labels: Core Roles, Adjuncts, Continuation Arguments and Co-referring Arguments, and (2) in each of the above class we apply the set of its specific classifiers, e.g. A0,..,A5 within the Core Role class. As such grouping is relatively new, the traditional features may not be sufficient to characterize each class. Thus, to generate a large set of features automatically, we again applied tree kernels.</Paragraph>
    <Paragraph position="4"> Since our SRL system exploits the PropBank formalism for internal data representation, we developed ad-hoc procedures to convert back and forth to the CoNLL Shared Task format. This conversion step gave us useful information about the amount and the nature of the parsing errors. Also, we could measure the frequency of the mismatches between syntax and role annotation.</Paragraph>
    <Paragraph position="5"> In the remainder of this paper, Section 2 describes the basic system configuration whereas Section 3 illustrates its enhanced properties and the hierarchical structure. Section 4 describes the experimental setting and the results. Finally, Section 5 summarizes  our conclusions.</Paragraph>
  </Section>
class="xml-element"></Paper>