A Framework for Artificial Intelligence.
Abstract
A framework is presented for several major ideas underlying Artificial Intelligence (AI) techniques. Two important components of this framework involve production systems and the predicate calculus. The thesis is proposed that a generalized production system architecture is the basic structure of most AI systems. Predicate calculus provides a language for the production rules and for the data structures on which these rules operate. Several different types of production systems are identified; these distinctions help to classify different kinds of AI systems. Two major types correspond roughly to AI systems for theorem proving and for robot problem solving. The last section of the report discusses similarities between a semantic network formalism and predicate calculus. The use of semantic networks in problems involving reasoning about class hierarchies and default information is described. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1979
- Accession Number
- ADA068188
Entities
People
- Nils J. Nilsson
Organizations
- SRI International