<?xml version="1.0" standalone="yes"?>
<Paper uid="H92-1095">
  <Title>Language Understanding Research at Paramax</Title>
  <Section position="3" start_page="0" end_page="0" type="metho">
    <SectionTitle>
Applications
2. SPOKEN LANGUAGE
UNDERSTANDING
</SectionTitle>
    <Paragraph position="0"> The focus of our spoken language understanding work is the development of a domain- and application- independent dialog processing architecture which can be coupled *This paper was partially supported by DARPA contract N000014-89-C0171, administered by the Office of Naval Research, and by internal funding from Paramax Systems Corporation (formerly Unisys Defense Systems).</Paragraph>
    <Paragraph position="1"> with a variety of speech recognizers. Some of our recent achievements include: * a non-monotonic reasoning capability, which will support &amp;quot;what if?&amp;quot; exploratory dialogs * a query paraphrase component to enhance user-friendliness null * enhanced reference resolution capabilities to support additional types of context-dependent references null * automated training techniques for semantics Figure 2 illustrates the integration of a speech recognizer, language understanding component, and a dialog manager (VFE) into a modular spoken language architecture interacting with the user and the application. A  more detailed description of this work can be found in \[1\].</Paragraph>
  </Section>
  <Section position="4" start_page="0" end_page="459" type="metho">
    <SectionTitle>
3. TEXT UNDERSTANDING
</SectionTitle>
    <Paragraph position="0"> In text understanding we emphasize an architecture in which a variety of components cooperate to produce  an analysis of texts. Not all components may be required for any one particular application - they can be intermixed in various ways to suit each application's needs. This architecture, shown in Figure 3 includes keyword-based information retrieval, for message routing. A knowledge-based information retrieval component (KBIRD) extracts text information which is accessable without a detailed natural language analysis. If an application requires detailed natural language processing, a natural language processing system is available. Finally, a database record generator formats the extracted information for database generation. This approach is described in \[2\].</Paragraph>
    <Paragraph position="1">  ~::~::~::~::~!~i:~ ................... ~:~% .................................... ::::::::i:i:i:i::&amp;quot;&amp;quot; &amp;quot;'::i:i &amp;quot; !:::i Mecqlng dettcr~l?ttoms of ~C/eat~ts -::!: ::::: ::::: :. -:: :::::|. ...... ~::::: ::::::::::::::::::::::::::::::::::::::i~:~:i~::.~i &amp;quot; ................ ~; ~:i~.::::~:::: ~mm~m~mm ~ I~!~i~:::::::::. !!!~i~i.':...~....:~iiii!i~i~i~i~ii! ::::::::::::::::::::::::::::::::::::::::: :: :::::::::::::: :: :::::::::: :: :::::::::::::::::::::::::::::::: :::: :: ::::::::::.&amp;quot;:~i~ii i~i YWORD-BASED INFORMATION R~TR1EVAL iF ..... * .... ,- lii ~ .: .__ ,, )</Paragraph>
  </Section>
  <Section position="5" start_page="459" end_page="459" type="metho">
    <SectionTitle>
4. DOCUMENT UNDERSTANDING
</SectionTitle>
    <Paragraph position="0"> The object of Paramax document understanding research is to develop novel knowledge-based approaches to the processing of document images. Starting with the scanned image(s) of a document's pages, our goal is to determine the document's functional and physical organization and to extract the key ideas from the ASCII representation of its text. Our current focus is on producing the structural interpretation of the document and the accompanying ASCII text for each component. The ASCII text would then be analyzed by a text-understanding system, as discussed in Section 2.</Paragraph>
    <Paragraph position="1"> An intelligent document understanding system is shown in Figure 4 with the scope of our project falling in the shaded region.</Paragraph>
    <Paragraph position="2"> This work is discussed in detail in \[3\].</Paragraph>
  </Section>
class="xml-element"></Paper>