Distributed Knowledge-Based Systems
Abstract
An introduction to and survey of recent research in the generic task approach to task-specific, knowledge-based systems is presented. Knowledge based systems using task specific architectures are steadily winning recognition as significant improvements over earlier attempts at building expert systems. There are generic problem solving strategies out of which complex knowledge based reasoning systems can be built. These generic strategies make use of knowledge organized in forms that are appropriate to the task and its characteristic control regime, and in that way help encode the expert's knowledge of a domain in a deeper way than is obtained using a typical rule based system. The research has developed the basic methodology of the generic task approach, has explored how deeper models and causal reasoning may be of use in problem solving systems, has helped in the production of a generic task toolset to encourage research on integration and useful in practical applications, and has examined ways in which the efficiency and robustness of systems can be improved-by learning, making use of parallelism, and by considering new environments for implementing generic tasks.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 15, 1989
- Accession Number
- ADA211984
Entities
People
- Balasubramanian Chandrasekaran
Organizations
- Ohio State University