Policy Teaching Through Reward Function Learning