Application of Deep Machine Learning Methods to Solve Mathematical Physics Problems in Manufacturing

This article discusses the importance of using deep machine learning methods to solve problems of mathematical physics in production. An overview of the main methods of deep machine learning - neural networks, convolutional and recurrent neural networks is provided. Forecasting of the technical condition of equipment, optimization of production processes, analysis of experimental data and modeling of physical processes are considered. Examples of successful application of deep machine learning methods in leading companies are given. An analysis of the advantages and limitations of using deep machine learning methods allows us to assess their effectiveness and possible problems when used in production.