Generative models aim to Density Estimation of dataGenerate new data from a posterior distribution that is as similar as possible to the distribution of training data. Then we can address better concern on unseen dataTutorial on Generative Adversarial Networks (2017)