The adversarial network was proposed by Goodfellow Ian in the paper Generative Adversarial Nets in 2014. In principle, adversarial networks can be simply summarized as a game process between a generator and a discriminator. During the whole network training process,
The division of labor between the two modules
- The judger, intuitively, is a simple neural network structure. The input is an image, and the output is a probability value, which is used to judge the true and false use (the probability value is greater than 0.5 is true, less than 0.5 is false)
- The generator can also be regarded as a neural network model, the input is a set of random numbers Z, and the output is an image. As can be seen from the figure, there will be two datasets, one is a real dataset, which is easy to say, and the other is a fake dataset, then this dataset is a dataset created by a generative network. Well, according to this picture, let's understand the goal of GAN is to do