File Information

File: 05-lr/acl_arc_1_sum/cleansed_text/xml_by_section/concl/89/h89-1013_concl.xml

Size: 2,060 bytes

Last Modified: 2025-10-06 13:56:22

<?xml version="1.0" standalone="yes"?>
<Paper uid="H89-1013">
  <Title>PORTABILITY IN THE JANUS NATURAL LANGUAGE INTERFACE 1</Title>
  <Section position="14" start_page="116" end_page="116" type="concl">
    <SectionTitle>
CONCLUSIONS
</SectionTitle>
    <Paragraph position="0"> KNACQ is based on the goal of allowing very rapid, inexpensive deFmition of a large percentage of the vocabulary necessary in a natural language interface to an expert system. It provides the knowledge engineer with the facilities to browse his/her taxonomic knowledge base, and to state head nouns, nominal compounds, and their non-clausal modifiers for referring to the concepts and roles in the knowledge base. Given that, KNACQ infers the necessary lexical syntactic and lexical semantic knowledge. Furthermore, if appropriate instances in the expert system knowledge base already have names, KNACQ will add proper nouns for those instances to the lexicon.</Paragraph>
    <Paragraph position="1"> KNACQ does not cover the inference of complex constructions typical of verbs and their nominalizations. IRACQ \[Ayuso, 1987\] allows a user to enter examples of usage for acquiring lexical syntax and semantics for complex constructions.</Paragraph>
    <Paragraph position="2"> Our experience thds far is that KNACQ has achieved our goals of dramatically reducing the time it takes to define the vocabulary for an expert system interface. It appears to have increased our own productivity several fold. (However, KNACQ has not yet been provided to a knowledge engineer with no knowledge of computational linguistics.) We believe that the problem of linguistic knowledge acquisition is critical not just as a practical issue regarding widespread availability of natural language interfaces. As our science, technology, and systems become more and more mature, the ante to show progress could involve more and more effort in filling domain-specific knowledge bases. The less effort spent on such knowledge bases, the more effort can be devoted to unsolved problems.</Paragraph>
  </Section>
class="xml-element"></Paper>
Download Original XML