In this repository I will upload the material of the course of Laboratory of Computational physics, of the Master degree in Physics of Data, during the academic year 2019/2020.
There is also present a folder "project" that contains the program realized for the final exam of this course, that consist into a complete and exhaustive study about Generative Adversarial Networks.
In particular, we focused on the following fields:
- brief study on the mathematical theory behind these networks, focusing in particular on the different games that can be used
- an algorithm designed to implement a DCGAN for the MNIST dataset
- a comparison between GANs and Variational Autoencoders
- a more complex algorithm for a DCGAN implemented for face generation