The idea is to let a DDQN learn how to play Mario via reinforcement learning.

I'm following the pytorch tutorial on how to build the network, what functions to add for RL, and how the algorithms are implemented.

In this I'm aiming to learn how to read papers then implement the content, familiarize myself with building classes for neural networks and building the network itself, and working with the pytorch environment.

Useful links:

• Source
• Overview of RL
• RL Cheatsheet
• Deep Reinforcement Learning with Double-Q-learning

Depending on when you read this most or all the links may be sourced from the pytorch website.