History

0.2.0 (2020-10-11)

• Added sparse Q-function implementation for tabular algorithms
• Made History objects JSON-serializable
• Added callbacks for logging and for capturing frames of a game
• Added preliminary support for Multi-agent Reinforcement Learning models
• Added BibTex citation in the README
• Misc bug fixing
• Improved tests

0.1.0 (2020-09-19)

• added support for basic agents: Q-Learning and SARSA;
• set up the development workflow of the project;
• added the main APIs, interfaces, abstract classes, to customize existing models or create new ones.
• shipped the first version of the documentation.