Towards Privacy-Preserving Machine Learning : Generative Modeling And Discriminative Analysis