Use of interacting neural networks in cryptography

Abstract

The computing power of computer technology, as well as the amount of data transmitted, is growing steadily every day. In this regard, existing encryption algorithms quickly become obsolete, due to insufficient cryptographic strength or bandwidth. There is a need to constantly create new encryption methods or modify old ones. As new algorithms can be algorithms using neural networks. This article presents the existing algorithms created on the basis of mutual synchronization of neural networks. The basic principles of their construction and application are described, as well as areas that require further research and modernization are identified.