APPROACH TO THE SYNTHESIS OF AN ALGORITHM PROCESSING ARTIFICIAL INTELLIGENCE DISTRIBUTIONS. Part 1

An innovative approach to the synthesis of an algorithm that processes artificial intelligence instructions is proposed. This approach hypothesizes that Markov networks with abnormal inference can serve as data sources for artificial intelligence. These networks are considered as a separate type of algorithm specifically for artificial intelligence, because they have a way of transmitting data to it and obtaining knowledge from it. This algorithm specifies a set of instructions that are triggered by the self-cocking mechanism of the Markov network with abnormal inference. These networks themselves are supposed to be built on the basis of Markov systems, consisting of сNnet-scheme of the Markov algorithm. The relationship of artificial intelligence with data sources for it and consumers of knowledge from it is explored. To implement the proposed approach, some terms related to the content of the conceptual idea of Markov networks with abnormal inference are clarified and generalized. Clarification of terms is a concretization of mathematical abstractions that can form the basis for further constructive discussions about methods for synthesizing artificial intelligence. Therefore, this work is theoretical in nature. Its goal is to prepare mathematical abstractions to begin the synthesis of constructive reasoning about artificial intelligence. The objective of the research is to constructively “link” the rather abstract concept of “artificial intelligence” to the objects of the world around us. These objects are considered primarily as data sources for artificial intelligence. Moreover, those interacting with it not directly, but through the mediation of Markov networks with abnormal inference, which transmit the results of processing data from objects of the world around us in the form of classified categories (topos) to artificial intelligence. At the same time, artificial intelligence displays its knowledge in the form of inverse morphisms to Markov networks and further from them to the objects of the world around us to make changes to their morphisms (possibly the objects of the world around us themselves).