I am working towards using RL to create an AI for a two-player, hidden-information, a turn-based board game. I have just finished David Silver's RL course and Denny Britz's coding exercises, and so am relatively familiar with MC control, SARSA, Q-learning, etc. However, the course was focused on single-player, perfect-information games, and I haven't managed to find any examples similar to the type of game I have, and would like advice on how to proceed.

I am still unsure how self-play works, and how it relates to MCTS. For example, I don't know if this involves using the latest agent to play both sides, or playing an agent against older versions, or training multiple opposing agents simultaneously. Are there good examples (or repositories) for learning self-play and MCTS for two-player games?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.