The Method of Mathematical Modeling of Energy Processes as a Mathematical Basis for Digital Twins of Neural Networks

This article is devoted to the development of the mathematical basis of digital twins of natural and artificial neural systems. Examples of such systems are: the nervous system of living organisms, analog (memristor) neurocomputers, chemical networks (chemical computers). The main advantage of analog microcomputers is the high speed of information processing, limited by the speed of analog components. Memristor components are the information storage of analog neurocomputers. However, for the effective implementation of analog neural systems, it is necessary to correctly define the structure and select the coefficients of such systems. To model the mentioned systems, it is proposed to use the method of mathematical prototyping of energy processes developed by the authors in the framework of mechanics, electrodynamics and modern nonequilibrium thermodynamics. The models constructed by this method are adequate, that is, they do not contradict the general physical laws and physics of a particular system, and they can also be arbitrarily accurate with enough experimental data. Therefore, in this paper, the method of mathematical prototyping of energy processes is considered as a mathematical basis for creating digital twins of natural and artificial neural systems.