Abstrakt: |
Artificial intelligence (AI) can be defined as a purview of computer science that automates reasonable human behavior (Luger George F., (2003) Stuart Russell and Peter Norvig (2020) ). The basis of reasonable human behavior is the decision of a person (Moiseev N.N., (1981)). Without a methodology for solving the tasks of the functioning of AI systems in the form of conditions for the existence of a process, we cannot guarantee the achievement of the goal of the activity. This situation gave rise to a fundamental problem: "The results of the activity in solving the tasks of functioning of AI systems do not meet the expectations of the decision maker (DM)." The decisions of the decision maker are based on three categories, such as: System - Model - Destination (Anokhin P.K., (1979), Arbib M.A., (1964), Burlov V.G., (2007)). Therefore, it is necessary to solve two tasks. Task 1. (System Development). Two approaches are known for system development (Goode H.H., Machol R.E., (1957)): development of a system based on analysis and development of a system based on synthesis. Synthesis is possible only on the knowledge of the law of construction and functioning of the developed system (Burlov V.G., (2018)). To solve problem 1 on the basis of synthesis, it is proposed to use the law of conservation integrity of the object (LCIO). LCIO guarantees the achievement of the goal of the functioning of the AI system. (LCIO is a stable, objective, repeating relationship between the properties of an object and the properties of its action with a fixed destination). Task 2. (Model development). The decision maker carries out activities based on the model. (Anokhin P.K., (1979), Arbib M.A., (1964), Burlov V.G., (2007)). To do this, you need to be able to synthesize adequate models. Achieving the goal of the functioning of artificial intelligence systems is possible only on the basis of a correctly constructed system and an adequate model. In well-known publications, this approach to the functioning of AI systems is absent. Therefore, the entire design of a correctly constructed system is implemented on the basis of the law of conservation integrity of the object, which confirms the advisability of considering the law of conservation integrity of the object as a condition for the existence of the functioning of AI systems. The model of the functioning of AI systems is based on the system integration of three processes: the process of problem formation, the process of identifying the problem, the process of neutralizing (eliminating) the problem. An indicator of the effectiveness of the functioning of AI systems is the likelihood that each problem is identified and neutralized. The simulation results confirmed the main trends in the process of functioning of the AI systems. [ABSTRACT FROM AUTHOR] |