Rusakova N. Design of brainlike structures and their application in artificial intelligence

The object of study is the process of a formal description of intellectual human activity on the basis of modeling brain-like structures. Research Methods - an algebra of finite predicates, algebra of predicate operations, graph theory, the basic concepts of category theory, mathematical tools of logical algebra and Boolean algebra. Equipment - a personal computer. Theoretical and practical research results - solution to an actual scientific problem of increasing the efficiency of information processing in artificial intelligence systems by developing methods and models for brain-like structures of human intellectual activity analysis using the algebra of finite predicates. The practical value of this work is to develop software products and hardware infrastructure, an experimental study of which confirms the effectiveness of the proposed approach. Scientific novelty - for the first time a method of exfoliation of the finite predicate has been proposed, which is characterized by a shift to a brain-like model structure, presented by a relational network that allows automated construction of relational networks. For the firsе time a model of a logical operator with manageable kernel, which is characterized by the introduction of the factors into the calculation of this operator and provides an opportunity to build a scheme of a separate branch of the relational network for variable sets. A method of minimizing formulas of Boolean algebra functions has been further developed for solving the problem of minimizing the number of states of the central pole of a relational network by introducing a bracket form of the predicate that led to a substantial reduction of the total number of poles' states of a relational network. Sound morphological model of the declination of nonpossessive adjectives has been improved, which in contrast to analogs does process potential word forms, it also takes into account the fluency of the last sound of base and it's built using the methods described in this paper, which enables the automated analysis of the grammatical structure of the text in the artificial intelligence systems. The results of this research are implemented at such companies as: Aldec Inc. (USA), "Special scientific researches" (Ukraine) and in the educational process of KhTURE (Ukraine). The scientific and practical results of the thesis can be used in research organizations and enterprises of Ukraine dealing with modeling human intellectual activity.