The Role of Databases in Knowledge-Based Systems.
Abstract
This paper explores the requirements for database techniques in the construction of knowledge-based systems (KBS). While early work in Artificial Intelligence (AI) has focused on techniques such as representation and problem-solving, scant attention was paid to the issues to which database (DB) research has focused (e.g., data sharing, query optimization, transaction processing). Our principal premise is that although it has appeared that there was little intersection between the particular focus of each group, there is a significant overlap in needs. The maturing of AI techniques has recently led to their application outside of the laboratory, thus thrusting upon them problems requiring DB solutions. On the other hand, DB needs have expanded to include more expressive data models and more powerful query languages (e.g., supporting inference). To ascertain KBS requirements for DBs, three KBSs are described. Each is analyzed from the perspective of the symbol and knowledge level concepts developed by Newell NEWE81. Limitations inherent in this perspective are identified. A new level, the organization level, is proposed as a means of identifying and dealing with these limitations. Lastly, we discuss implementations of some of the requirements in the SRL knowledge engineering system. Three knowledge-based systems are reviewed: XCON/R1, ISIS, and Callisto.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1986
- Accession Number
- ADA166365
Entities
People
- John Mcdermott
- Mark S. Fox
Organizations
- Carnegie Mellon University