Reinforcement learning in Minecraft: Challenges and opportunities in multiplayer games
During this session you will explore: The Malmo platform and multi-agent tasks Using the reinforcement learning library RLlib to implement and train agents to complete Minecraft tasks Coordinated policies for collaborative multi-agent tasks Open challenges in learning robust policies for ad-hoc teamwork
