<?xml version="1.0" standalone="yes"?>
<Paper uid="W02-0227">
  <Title>A Minimum Message Length Approach for Argument Interpretation</Title>
  <Section position="8" start_page="0" end_page="0" type="concl">
    <SectionTitle>
6 Conclusion
</SectionTitle>
    <Paragraph position="0"> We have offered a mechanism which produces interpretations of segmented NL arguments. Our application of the MML principle enables our system to handle noisy conditions in terms of wording, beliefs and argument structure, and allows us to isolate</Paragraph>
    <Paragraph position="2"> the effect of the underlying knowledge representation on the interpretation process. The results of our automated evaluation were encouraging, with interpretations that match perfectly or almost-perfectly the source-BN being generated in 75% of the cases under all noise conditions.</Paragraph>
    <Paragraph position="3"> Our system has the following limitations: a1 The interpretations generated by our system are in terms of the propositions and relations known by the system. However, the MML Principle itself addresses this limitation (at least partially), as the length of a message is a quantitative measure for determining whether an interpretation is likely to reflect the user's intentions.</Paragraph>
    <Paragraph position="4"> a1 Our mechanism does not infer an implicit goal proposition, nor does it infer discourse relations from free-form discourse. At present, this limitation is circumvented by forcing the user to state the goal proposition of the argument, and to indicate clearly the antecedents and consequents of the implications in his/her argument (this is done by means of a web-based interface).</Paragraph>
    <Paragraph position="5"> a1 Our argument-interpretation mechanism has been tested on one knowledge representation only - BNs.</Paragraph>
    <Paragraph position="6"> a1 It is unclear whether arguments produced by automatically distorting our system's arguments are representative of arguments generated by people.</Paragraph>
    <Paragraph position="7"> Further trials with real users will be conducted to ascertain this fact.</Paragraph>
    <Paragraph position="8"> a1 The system's performance deteriorates for large BNs (9 nodes). However, it is unclear whether this will affect the use of the system in practice.</Paragraph>
    <Paragraph position="9"> Despite these limitations, we are hopeful about the potential of this approach to address the discourse interpretation challenge.</Paragraph>
  </Section>
class="xml-element"></Paper>