Problems of Constructing Intelligent Systems. Knowledge Representation

  • PDF / 158,572 Bytes
  • 10 Pages / 594 x 792 pts Page_size
  • 16 Downloads / 210 Views

DOWNLOAD

REPORT


CYBERNETICS PROBLEMS OF CONSTRUCTING INTELLIGENT SYSTEMS. KNOWLEDGE REPRESENTATION

V. Yu. Meytus

UDC 004.896

Abstract. This article considers the construction of a formal model of a subject area for an intelligent system using intelligence with descriptive logic. Three interrelated category models are created. In particular, the representation of knowledge about a subject area is defined as a category of knowledge. Some properties of the introduced categories and the relationship between the category of knowledge and the ability to solve problems in the subject area being modeled are considered. Keywords: intelligence, intelligent system, subject area, description logic, subject area model, knowledge, intensional, ontology, category of knowledge. INTRODUCTION This article is a continuation and development of ideas related to the construction of intelligent systems (ISs) and described in [1] and uses concepts and definitions similar to those from [1]. The main problem considered in this article is the development of means and methods for representing information for ISs about a subject area (SbA) in the form of knowledge oriented to solving problems in this SbA. By the intelligence of a subject we understand the subject property that allows him to adequately model the SbA that is perceived by him and with which he interacts and to solve problems of interaction between him and the SbA at the level of the model constructed. An intelligent system is understood to be a subject possessing intelligence, acting in an SbA, and using his intelligence to organize interaction with the SbA. Note that, depending on modeling methods, the existence of different SbA models is possible, for example, models for representing the outside world from the Newton physics to the Einstein physics, from quantum physics to string theory, from the Minkowski four-dimensional model of the world or the 10-dimensional string theory to the Shapiro–Virasoro model with 26 dimensions of space-time. Different models are used in the classical and constructive mathematics or in groups of permutations and ð-adic numbers in group theory. In fact, mathematics is a science about the formation of behavior (of sequential problem solving) in various formal models. And the intelligence of a mathematician allows him to construct a model and, based on certain logic, to solve problems in it. The ontological SbA model proposed in [1] actually includes many different models for the same SbA that are distinguished from one another, first of all, by ontologies and logics related to them. The main advantage of these models is their connection with natural language, which provides the understanding of the model constructed and processes of solving problems by the user with its help. However, in the general case, by its parameters, properties, and components, an SbA is wider than an ontological model since, mostly, not all SbA elements are presented in its ontology. Therefore, when proving certain results, ontological definitions of new SbA components are formed that ca