arXiv 2506.00458
Reinforcement Learning for Hanabi
By Nina Cohen and Kordel K. France
Published 2025-05-31
Discussion
Read the public discussion and references gathered around this paper.
Hanabi has become a popular game for research when it comes to reinforcement learning (RL) as it is one of the few cooperative card games where you have incomplete knowledge of the entire environment, thus presenting a challenge for a RL agent. We explored different tabular and deep reinforcement learning algorithms to see which had the best performance both against an agent of the same type and also against other t…