This blog is about my project ‘GANArt: Using AI to create Portraits,’ for which I have used various public domain portraits between 1750 to 1950 to train the AI model. The model then adapts the art style and then generates new unique portraits on its own. GAN(Generative Adversarial Network) is no wonder an amazing discovery in the field of Deep Learning. From generating fake human faces to an entirely new virtual world, if there is room for creativity, one can indeed implement GAN.

“Everything you can imagine is real." Pablo Picasso Introduction to Generative ModelingTo explain Generative Modeling in layman’s terms lets think of a scenario where our model is a classifier and given an image the model classifies it into a cat or dog image, now what if we want this process to go in the opposite direction where we provide a prescription of what we want to our model and model generates an image corresponding to that prescription, that is a generative model in its simplest form.