ganIsing

This is my final project for STAT241 / CS281A: Statistical Learning Theory at UC Berkeley in Fall 2021.

Given that I’ve gotten out of Physics just about when the Deep Learning “revolution” hit the field, it was a great reason to try in person how generative adversarial networks can replicate results obtained by time-consuming Monte Carlo simulation, something I am deeply familiar with.

The goal of this project was two-fold: to examine the approach on a classic example - the 2D Ising model, and to provide a brief (the final report linked above was limited to 10 pages) overview of the use of GANs in general.