The Application of Neural Networks in Cryptography

In this paper, it will be shown cooperation between Petri nets and neural networks. The approach to the formation of the structure of a neural network with the help of Petri nets model , which you can describe the algorithms and, in particular, encryption algorithms Built model in networks Petri, according to the proposed approach, is the basis for further construction neural network. The idea of an informal transformation, which makes sense from because the structure of the Petri net provides a justification for the structure of the neural network, which leads to a decrease in the number of parameters for training in the neural network (in the considered). At the same time, training is only a fine adjustment of the parameter values. Also the transformation can be explained by the fact that both Petri nets and neural networks are function description languages, with the difference that in the case of neural networks, the function being set must first be trained (or find the values of the parameters). This is discussed by the example of a simulator for encryption algorithms nets. Also, in the article Probabilistic Petri Nets and their properties are described, except for the technique of their use for system's research. 5 theorems are proved and 5 generalizing conclusions about application of languages of to a problem of Probabilistic Petri nets resolvability are made.