KASH: A General Purpose Knowledge Acquisition Shell
Abstract
The field of expert systems has numerous and successful applications ranging from simple tasks to very complex ones. To construct a system requires the developer to isolate a set of rules that will guide the decision making process. Rules are conventionally defined during a series of interviews between a domain expert and a knowledge engineer. The encoding and representation of the extracted domain knowledge has proven to be a difficult barrier to overcome. KASH (Knowledge Acquisition Shell) has been designed to address this problem by providing a knowledge engineer with a set of utilities for constructing knowledge acquisition sessions base on interviewing techniques. The information elicited from domain experts during the sessions is guided by a question dependency graph (QDG). The QDG, defined by the knowledge engineer, is a series of control questions about the domain that are used to organize the cognitive processes of an expert. The content information supplied by the expert, in response to the questions, is represented in the form of a concept map. These maps can be constructed in a top-down or bottom-up manner by the QDG and used by KASH to generate the rules for a large class of expert domains. Expert systems, Artificial intelligence, Knowledge acquisition knowledge engineer, Domain expert, KASH, QDG, Rules, Knowledge base.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1990
- Accession Number
- ADA221553
Entities
People
- Christopher R. Westphal
- Duc T. Tran
Organizations
- Assistant Secretary of the Air Force