<?xml version="1.0" standalone="yes"?>
<Paper uid="A88-1018">
  <Title>INTEGRATING TOP-DOWN AND BOTTOM-UP STRATEGIES IN A TEXT PROCESSING SYSTEM</Title>
  <Section position="3" start_page="0" end_page="0" type="intro">
    <SectionTitle>
INTRODUCTION
</SectionTitle>
    <Paragraph position="0"> The System for Conceptual Information Summarization, Organization and Retrieval (SCISOR) is an implemented system designed to extract information from naturally occurring texts in constrained domains. The derived information is stored in a conceptual knowledge base and retrieved using a natural language analyzer and generator.</Paragraph>
    <Paragraph position="1"> Conceptual information extracted from texts has a number of advantages over other information-retrieval techniques \[Rau, 1987a\], in addition to allowing for the automatic generation of databases from texts.</Paragraph>
    <Paragraph position="2"> The integration of top-down, expectation driven processing, and bottom-up, language-driven parsing is important for text understanding. Bottom-up strategies identify surface linguistic relations in the input and produce conceptual structures from these relations. With the input &amp;quot;ACE made ACME an offer&amp;quot;, a good &amp;quot;bottom-up&amp;quot; linguistic analyzer can identify the subject, verb, direct and indirect objects. It also can determine that ACME was the recipient of an offer, rather than being made into an offer, as in &amp;quot;ACE made ACME a subsidiary&amp;quot;.</Paragraph>
    <Paragraph position="3"> Top-down methods use extensive knowledge of the context of the input, practical constraints, and conceptual expectations based on previous events to fit new information into an existing framework. A good &amp;quot;top-down&amp;quot; analyzer might determine from &amp;quot;ACE made ACME an offer&amp;quot; that ACME is the target of a takover (which is not obvious from the language, since the offer could be for something that ACME owns), and relate the offer to other events (previous rumors or competing offers).</Paragraph>
    <Paragraph position="4"> Bottom-up methods tend to produce more accurate parses and semantic interpretations, account for subtleties in linguistic expression, and detect inconsistencies and lexical gaps. Top-down methods are more tolerant of unknown words or grammatical lapses, but are also more apt to derive erroneous interpretations, fail to detect inconsistencies between what is said and how it is interpreted, and often cannot produce any results when the text presents unusual or unexpected information. Integration of these two approaches can improve the depth and accuracy of the understanding process.</Paragraph>
    <Paragraph position="5"> SCISOR is unique in its integration of the bottom-up processing performed by its analyzer, TRUMP (TRansportable Understanding Mechanism Package) \[Jacobs, 1986\], with other sources of information in the form of conceptual expectations.</Paragraph>
    <Paragraph position="6"> In this paper, four information sources are described that are used by SCISOR to produce meaning representations from texts. The actual processing sequence and timing of the application of these sources are illustrated.</Paragraph>
  </Section>
class="xml-element"></Paper>