Abstract: Autonomous vehicles trained through multiagent reinforcement learning (MARL) have shown impressive results in many driving scenarios. However, the performance of these trained policies can ...