<?xml version="1.0" standalone="yes"?>
<Paper uid="C92-3167">
  <Title>Recognizing Topics through the Use of Interaction Structures</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> An aggregation of sentences having local coherence is called a &amp;quot;discourse segment&amp;quot;. Such a structure must be recognized to understand discourse including dialogues. The structure constrains candidates, for example, referents for anaphora resolution and plans for plan recognition. A topic is a kind of local coherence. Segments can be recognized in both task-oriented and non-task-oriented dialogues because most dialogues have explicit topics.</Paragraph>
    <Paragraph position="1"> Recognized topics can also be used in a topic-oriented video retrieval snpport system. The system recognizes the topic structures of video sequences such as documentaries, and shows a topic list. Topic nests are expressed by indentation. Users can survey the contents of a video library, and play back sequences connected to an interesting topic.</Paragraph>
    <Paragraph position="2"> This paper describes how to recognize topics of both task-oriented and non-task-oriented dialogues without domain knowledge. First, a basic topic recognition mechanism is discussed.</Paragraph>
    <Paragraph position="3"> Second, identifying topic continuation through the interaction structure is presented. Finally, coverage of the interaction structure approach is discussed.</Paragraph>
  </Section>
class="xml-element"></Paper>