The framework for measuring productive models in a competitive process in which we simultaneously train two models: a productive model G. which captures the distribution of the data and also a biased model D. The probability that a sample appeared in the training data than G. The training process G is supposed to increase the chances, that D will make a mistake. This slide is accompanied by a small game for two players. In the artificial space problems G and D, there is a unique solution where G accepts the training data distribution and D is equal to 1 2 everywhere. In case G and D are defined using multilayer perceptron’s, the whole system can be trained using backpropagation. To build a well-performing generator, which makes use of machine learning algorithms to produce the required outputs.Solve this using neural network to Generate a Photo Realistic Images using GAN (Generative Adversarial Networks). To gain new information from the generated images. The main idea of generative adversarial network can be compared to game of two players - here two players are generator and discriminator framework for measuring productive models in a competitive process in which we simultaneously train two models: a productive model G. which captures the distribution of the data and also a biased model D. The probability that a sample appeared in the training data than G. The training process G is supposed to increase the chances, that D will make a mistake. This slide is accompanied by a small game for two players. In the artificial space problems G and D, there is a unique solution where G accepts the training data distribution and D is equal to 1 2 everywhere. In case G and D are defined using multilayer perceptron’s, the whole system can be trained using backpropagation

.