Use of the knowledge which is independence on reward in reinforcement learning | IEEE Conference Publication | IEEE Xplore