<?xml version="1.0" standalone="yes"?>
<Paper uid="M93-1024">
  <Title>Tagged Sentences Tagger Lisp-readable sentences CompletedChart Template Generator, Target Templates Input Text System Knowledge Bases</Title>
  <Section position="1" start_page="0" end_page="0" type="abstr">
    <SectionTitle>
E-mail: lytinen Oengin.umich.edu
BACKGROUND
</SectionTitle>
    <Paragraph position="0"> Over the past five years, we have developed a natural language processing (NLP) syste m called LINK . LINK is a unification-based system, in which all syntactic and semantic analysi s is performed in a single step . Syntactic and semantic information are both represented in th e grammar in a uniform manner, similar to HPSG (Pollard and Sag, 1987) .</Paragraph>
    <Paragraph position="1"> LINK has been used in several information extraction applications . In a project with General Motors, LINK was used to process terse free-form descriptions of symptoms displayed b y malfunctioning automobiles, and the repairs which fixed them . In this very narrow domain , LINK achieved recall and precision rates of 80-85% .</Paragraph>
    <Paragraph position="2"> Most recently, we used the LINK system to participate in MUC-4. During this competition , we developed initial versions of pre and postprocessing modules which were further developed in MUC-5. In MUC-4, LINK achieved recall and precision rates of about 40% .</Paragraph>
  </Section>
class="xml-element"></Paper>