<?xml version="1.0" standalone="yes"?>
<Paper uid="P05-1072">
  <Title>Semantic Role Labeling Using Different Syntactic Views[?]</Title>
  <Section position="18" start_page="587" end_page="587" type="concl">
    <SectionTitle>
7 Conclusions
</SectionTitle>
    <Paragraph position="0"> We described a state-of-the-art baseline semantic role labeling system based on Support Vector Machine classifiers. Experiments were conducted to evaluate three types of improvements to the system: i) adding new features including features extracted from a Combinatory Categorial Grammar parse, ii) performing feature selection and calibration and iii) combining parses obtained from semantic parsers trained using different syntactic views. We combined semantic parses from a Minipar syntactic parse and from a chunked syntactic representation with our original baseline system which was based on Charniak parses. The belief was that semantic parses based on different syntactic views would make different errors and that the combination would be complimentary. A simple combination of these representations did lead to improved performance.</Paragraph>
  </Section>
class="xml-element"></Paper>