Generative Adversary Networks in Signal Generation Problems

The object of the research is a multilayer perceptron which is being considered as a base neural network for the generative adversary neural network. The goal of the study is to make a neural network and make observations about how it can possibly work on a problem with separating: linear separable, semi separable and not separable subsets. The methodology of the study is a representation of the mathematical model of the multilayer perceptron on a python programming language. The results of the study are represented on the figures 2 - 5. The future application of the results are relevant for the further analysis in high particle physics for the track identification problem. Conclusions on the research are mostly based on the results with the tasks of separating objects, the accuracy of the dividing subsets is important for the future basis analysis on the subsets.