PROCESSING DIFFERENT INFORMATION USING THE DEEP LEARNING OF NEURAL NETWORKS

UDC 004.032.26

Authors

A.V. Potemkin

Abstract

The paper proposes the architecture of a neural network for processing heterogeneous information. The proposed architecture includes separate inputs for text, graphic, tabular, graph and meta information. An experimental study of the performance indicators of the developed architecture was carried out to solve the problem of identifying information operations in the social network Internet.

Keywords

heterogeneous information, neural networks, deep learning, universal architecture.