How to Build GANs To Synthesize Data

If you're working in deep learning, you've probably heard of GANs, or Generative Adversarial Networks (Goodfellow et al, 2014). In post we will explain what GANs are and discuss some use cases with examples. I am adding to this post a link to my GAN playground, called MP-GAN (Multi Purpose GAN). I prepared this playground in github as a research framework, and you are welcome to use it to train and explore GANs for yourselves. GANs are part of a family of generative deep learning architectures, whose goal is to generate synthetic data, instead of predicting features of data points (these are the more common discriminative models, such as classifiers and regressors.