Abstract: Reinforcement learning (RL) has achieved promising results in solving numerous challenging sequential decision problems. To address the issue of sparse extrinsic rewards, researchers have ...