<?xml version="1.0" standalone="yes"?>
<Paper uid="W98-0202">
  <Title>Intelligent Network News Reader with Visual User Interface</Title>
  <Section position="2" start_page="0" end_page="12" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> Extracting necessary information easily from the bulk of information available throughout the world is crucial for people living in this highly computerized society, and therefore, it is necessary to develop systems which can visually present the selected information necessary to assist people in forming new concepts. A great deal of work on this subject has been done by various researchers, e.g., information retrieval from newspaper articles and message understanding in newspaper articles.</Paragraph>
    <Paragraph position="1"> It is not sufficient that this kind of expert system simply imitate the real world. Such systems have to create a richer environment with the visual interface than there is now. This means not simply supplying an imitation of the real world, but actively building a virtual world where the density of information is higher than that of the real world for a given use. In other words, we need information screening for each individual user. Therefore, technology which selects and presents the necessary information will be the key to information retrieval in the future.</Paragraph>
    <Paragraph position="2"> As an application of this kind of technology to the Internet, we are developing an Intelligent Network News Reader (HISHO: Helpful Information Se- null lection by Hunting On-line) which extracts news articles for users and which visually displays the structure of articles. In contrast to ordinary information retrieval and abstract generation, this method utilizes an &amp;quot;information context&amp;quot; to select articles from newsgroups on the Internet. We finished our prototype of the Intelligent Network News Reader in March 1998 and will complete a final practical version in March 2000.</Paragraph>
    <Paragraph position="3"> In this paper, we discuss how to find topic changing articles in the tree structures of news articles, how to extract topic differences from the thread of articles, and how to indicate this information in the display to help users decide which part of the tree structures of articles they will read.</Paragraph>
  </Section>
class="xml-element"></Paper>