Features of the Neural Network for Determining the Position and Geometric Characteristics of Cavernous Inclusions

This article describes the process of developing a neural network (NN) capable of restoring the structure of the geological and physical model of the medium (GPMM) based on a known picture of the propagation of a wave field. The architecture of the NN itself and its components are indicated, information on the process of its training is provided. Also, the time of operation of the NN on different devices and the results obtained are shown. The work was supported by the Novosibirsk State Technical University (Project C-19, 2018).