Abstract: Multi-agent Reinforcement Learning (MARL) algorithms have shown remarkable success in tasks requiring cooperation. However, in sparse reward scenarios, traditional MARL algorithms struggle ...