Home

This Repo is all about my journey in to Reinforcement Learning (hence the name). Right now it is totally lacking in any form of organization.

For the Roadmap:

• Change layout so that each set of environments (provided that there is something which can be realistically called a set) is contained in its own folder inside the Master. Inside those, projects should be sorted by algorithm type used.

• For the deep Q learning Atari 2600 part of this repository I plan on implementing replay memory prioritization and noisy networks for exploration, in addition to the already implemented replay buffer, double q learning, and a dueling network architecture.