Automated Acquisition of Evolving Informal Descriptions
Abstract
Much of human communication proceeds via an exchange of informal descriptions characterized by ambiguity, contradiction, and incompleteness. This thesis describes an automated system, the Listener, that unintrusively performs knowledge acquisition and informal input. The Listener develops a coherent internal representation of a description from an initial set of disorganized imprecise statements. Each statement provides fragmentary, ambiguous, possibly inconsistent information. The description and its implications are checked for consistency, completeness, and conclusions deemed interesting. The Listener interactively presents the results of these checks as a guide to continuing the acquisition process. The Listener also produces a summary document from its internal representation in order to facilitate communication, review, and validation. These behaviors are supported by a variety of techniques, including dependency-directed reasoning, hybrid knowledge representation, and the reuse of common forms (cliches). (Author) (kr)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1990
- Accession Number
- ADA225621
Entities
People
- Howard B. Rubenstein
Organizations
- Massachusetts Institute of Technology