Distributed Knowledge Base Systems for Diagnosis and Information Retrieval.
Abstract
During the year, progress was made in our research on distributed approaches to knowledge-based problem-solving in the following areas: 1) We have developed an approach called consolidation to reason qualitatively about the behavior of systems of components. This approach and the more classical approach of qualitative simulation are complementary. 2) We have made further progress in developing an architectural framework for diagnostic reasoning. 3) We have elucidated some of the criteria that govern how design plans are selected for further refinement in design problem solving. 4) We have identified a number of generic tasks into which the information processing activity of most of the expert systems can be decomposed. These generic tasks are at a much higher level of abstraction, and this should make knowledge acquisition and explanation for expert systems easier. 5) We have clarified how symbolic qualitative knowledge-based processing helps when problems get complex by considering the concrete task of classification. We compare the pattern recognition and AI approaches to the problem. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1985
- Accession Number
- ADA170830
Entities
People
- Balasubramanian Chandrasekaran
Organizations
- Ohio State University