This project was for CIS 519, Applications of Machine Learning and I worked on it with Lakshay Sharma and Sean Waxmen-Lenz. The goal of our project was to utilize techniques learned in class such as reinforcement learning and neural networks to develop a machine learning algorithm that outputs an AI to play Super Mario Bros.

To develop our algorithm, we used OpenAI Gym for our environment and Google colab to work concurrently as a group. We focused on reinforcement learning for our approach and specifically utilized an epsilon greedy Q learner model. The main steps to the project included setting up the environment, constructing a neural network for image processing, developing our model, experimenting with reward systems, and testing hyperparameters.