A new approach for structural credit assignment in distributed reinforcement learning systems | IEEE Conference Publication | IEEE Xplore