Accession Number:

ADA222292

Title:

Semi-Automatic Methods of Knowledge Enhancement

Descriptive Note:

Final rept. Sep 1986-Dec 1988

Corporate Author:

TURING INST GLASGOW (UNITED KINGDOM)

Personal Author(s):

Report Date:

1988-12-05

Pagination or Media Count:

139.0

Abstract:

The objectives of the project are concerned with the representation of human concepts in machines. The issues addressed are 1 How to automate the development of concepts from examples supplied by an expert. 2 How to automate the validation of factual databases which have been generated by a machine. 3 How to automate the development of concepts summarising the information in comprehensive databases of raw facts. 4 How to achieve 3 above for problem domains so complex that no effective human concepts exist, or can be constructed by expert brains Ultra-Complex Domains or UCDs. 5 How to develop machine methods for imparting these concepts to human experts, thus providing a method for mastering what was previously impenetrable. For reasons of economy and ease of control, the chose problem domains were chess endgames. Substantial results have been achieved with automatic structuring of problems and the design of primitive starting concepts, previously the province of the human intellect. The investigators believe that this work has forced an entrance to territory new to computer science.

Subject Categories:

  • Human Factors Engineering and Man Machine Systems

Distribution Statement:

APPROVED FOR PUBLIC RELEASE