Towards Skill and Population Curriculum for MARL

Rundong Wang, Longtao Zheng, Wei Qiu, Bowei He, Bo An, Zinovi Rabinovich, Yujing Hu, Yingfeng Chen, Tangjie Lv, Changjie Fan

Keywords: reinforcement learning, mutlit-agent system, curriculum learning, multi-armed bandit

Abstract: Recent advances in multi-agent reinforcement learning (MARL) allow agents to coordinate their behaviors in complex environments. However, common MARL algorithms still suffer from scalability and sparse reward issues. One promising approach to resolve them is automated curriculum learning (ACL), where a student (curriculum learner) train on tasks of increasing difficulty controlled by a teacher (curriculum generator). Unfortunately, in spite of its success, ACL's applicability is restricted due to: (1) lack of a general student framework to deal with the varying number of agents across tasks and the sparse reward problem, and (2) the non-stationarity in the teacher's task due to the ever-changing student strategies. As a remedy for ACL, we introduce a novel automatic curriculum learning framework, Curriculum Oriented Skills and Tactics (COST), adapting curriculum learning to multi-agent coordination. To be specific, we endow the student with population-invariant communication and a hierarchical skill set. Thus, the student can learn cooperation and behavior skills from distinct tasks with a varying number of agents. In addition, we model the teacher as a contextual bandit conditioned by student policies. As a result, a team of agents can change its size while retaining previously acquired skills. We also analyze the inherent non-stationarity of this multi-agent automatic curriculum teaching problem, and provide a corresponding regret bound. Empirical results show that our method improves scalability, sample efficiency, and generalization in MPE and Google Research Football. The source code and the video can be found at https://sites.google.com/view/neurips2022-cost/.