From Few to More: Large-scale Dynamic Multiagent Curriculum Learning

Published in Proceedings of AAAI Conference on Artificial Intelligence (AAAI), 2020


This paper proposes a novel Dynamic Multiagent Curriculum Learning (DyMA-CL) to solve large-scale problems by starting from learning on a multiagent scenario with a small size and progressively increasing the number of agents. We propose three transfer mechanisms across curricula to accelerate the learning process. Experimental results show that DyMA-CL using DyAN greatly improves the performance of large-scale multiagent learning compared with state-of-the-art deep reinforcement learning approaches.