<?xml version="1.0" standalone="yes"?>
<Paper uid="W05-0639">
  <Title>The Integration of Syntactic Parsing and Semantic Role Labeling</Title>
  <Section position="5" start_page="238" end_page="239" type="evalu">
    <SectionTitle>
5 Results and Discussion
</SectionTitle>
    <Paragraph position="0"> Table 1 shows the results after combining various SRL systems using different parsers. In order to explore the effects of combining, we include the over-all performance on the development dataset of individual SRL systems in Table 2.</Paragraph>
    <Paragraph position="1"> The performance of Semantic Role Labeling (SRL) is determined by the quality of the syntactic information provided to the system. In this paper, we investigate that for the SRL task whether it is more suitable to use a parser trained with data con- null the WSJ test (bottom).</Paragraph>
    <Paragraph position="2"> taining both syntactic bracketing and semantic argument boundary information than a pure syntactic one.</Paragraph>
    <Paragraph position="3"> The results of the SRL systems using the AMor AN- parsers are not significantly better than that using the Charniak's parser. This might due to the simple training mechanism of the base parsing algorithm which the AM- and AN- parsers exploit. It also suggests our future work to apply the approach to more sophisticated parsing frameworks. By then, We show that we can boost the final performance by combining different SRL systems using different parsers, given that the combination algorithm is ca- null individual SRL systems.</Paragraph>
    <Paragraph position="4"> pable of maintaining the quality of the final arguments. null</Paragraph>
  </Section>
class="xml-element"></Paper>