A dualism of neural networks

Abstract

Neuroinformatic systems differ from traditional informatics systems in the dual nature of their information processing, namely, the dualism of neural networks. It is proposed to use analogy criteria of neural network dualism properties for comparing natural and artificial neural networks. The use of such analogy criteria makes it possible to obtain a high degree of correspondence between the structure and function characteristics of artificial and real neural networks when simulating neuroinformatic systems. >