KRS(ID:2147/krs001)

Knowledge representation system 

for the Knowledge Representation System

van Marke VUB Belgium 1987

Frame-based language built on Common LISP.
Related languages
Common LISP => KRS   Built on
References:
  • Van Marcke, Kris "A Consistence Maintenance System based on Forward Propagation of Proposition Denials" VUB AI-Lab TR 86-01 view details Abstract: FPPD is a Consistency Maintenance System based on forward propagation of proposition denials. It was designed to support an automatic caching-management system for the knowledge representation system KRS and to guarantee integrity in a non-monotonically changing knowledge base. FPPD can be used independently to support problem solving which involves there consideration of assumptions. Two assumptions underly the design. The application must rely a lot on dependency-constraints and the frequency of questioning values must be notably higher than the frequency of changing them.The fundamental difference with earlier Truth Maintenance Systems is that the user does not define the dependency-relations explicitly. FPPD is continuously shifting parts of the dependency-network from the implicit description (as it is defined by the user) to explicit dependency-relations and vice-versa.
  • Strickx, Peter "Extrak: A Second Generation Expert System in KRS" VUB AI-Lab TR 87-07 view details Abstract: First generation expert systems are too weak to solve large and extending problems. The rules fail abruptly when unanticipated situations are encountered. Second generation expert systems solve these problems by using a deep model to gradually generate their shallow knowledge. Extrak is a second generation expert system tool because it provides a framework for experiments with the combination of shallow and deep knowledge. The system uses shallow knowledge when available and otherwise falls back on its deep model of the problem domain. Meta-rules guide the reasoning process. Factual and control knowledge are represented uniformly in KRS (the knowledge representation system).
  • Van Marcke, Kris "Context Determination through Inheritance in KRS" VUB-AI-Lab TR 87-01 1987 view details Abstract: KRS is a knowledge representation system developed at the VUB-AI-Lab in Brussels. It supports the use and collaboration of multiple formalisms. A notion of meaning, inspired by intensional logic, provides a flexible framework for representation. The KRS concept-language is based on a single inheritance tree. We argue that a single inheritance hierarchy contributes to cleaner representational structures. We also provide a mechanism for explicit and proper redirection of inheritance where needed. A context retrieval mechanism, based on the lexical scope rule and determined by the inheritance tree, provides implicit information about how a concept fits into a larger cluster of concepts.
  • Van Marcke, Kris "The KRS Manual" VUB AI-Lab TR 87-03 view details
  • Rademakers, ilip "Implementing Reflective Architectures in Object Oriented Languages" VUB AI-Lab TR 88-14 view details Abstract: Computational reflection is the activity performed by a computational system when reasoning about aspects of itself. In recent years several authors have integrated reflection in procedural languages such as \s-2LISP and PROLOG. However, less work has been done for Object-oriented languages. This thesis focuses on the construction of a reflective architecture for a prototype-based object-oriented knowledge representation language KRS. The design of the architecture is discussed and its implementation is presented in great detail. The second part of this work elaborates on reflection in class-based object-oriented languages. More particularly it shows how a clean implementation of the meta-class concept leads to a powerful and extensible language.
  • Van Marcke, Kris "BKRS: An Object-Oriented Representation Language" VUB-Lab TR 88-09 view details Abstract: This paper describes the representation language KRS. KRS is a language developed at the VUB AI-Lab, inspired by representation language languages and intensional logic. Its main contributions are the representation axiom and the caching and consistency maintenance mechanisms. The representation axiom is a constraint, defining the relation between referents and definitions. It combines a denotational and a computational style of representation. Caching is a mechanism to remember computed values for future re-use. A consistency maintenance system keeps cached values conform with their definitions.
  • Van Marcke, Kris "KRS User Manual" VUB AI-Lab TR 88-15 view details Abstract: This document contains the KRS user manual, describing a Symbolics Common-Lisp implementation of KRS release 3.0.
  • Steels, Luc & Van Marcke, Kris "The KRS Tutorial" VUB AI-Lab TR 89-16 view details Abstract: This document contains a tutorial of the knowledge representation system KRS.It supplements the KRS manual which should be used for further reference.
  • Gaines, Brian R. "Empirical investigation of knowledge representation servers: design issues and applications experience with KRS" view details
          in ACM SIGART Bulletin 2(3) June 1991 Special issue on implemented knowledge representation and reasoning systems view details