ARTIFICIAL INTELLIGENCE – DIGITAL TECHNOLOGIES AND COGNITIVE FUNCTIONS

is that use the latest achievements of human science in the field of integrating cognitive functions into digital technologies. Modern achievements of domestic and foreign researchers in the fields of neurobiology and information technology are presented. The basic elements of artificial neural networks are described in detail. The key limitations of existing models of artificial neural networks are analyzed, such as narrow specialization, lack of sustainable opportunities for implementing the mechanism of emotionality, insufficient flexibility and adaptability to changes in the conditions of the environment of their functioning. The authors provide an overview of classical and modern architectures and methods of training neural network models, and also discuss the problems of their application, indicating how to increase the efficiency of their use. The work pays special attention to the need for indepth research in the field of creating new learning algorithms, increasing computing power, as well as in the field of neurobiology. The problems of existing definitions of artificial intelligence are described from the point of view of their universality. The authors propose a method for increasing the efficiency of neural networks, based on their disadvantages and advantages. The article will be of interest to specialists in the field of artificial intelligence, researchers in the field of integration of cognitive functions into digital technologies.