reinforcement learning is the only practical way to do learning because it can be hard to find labeled training examples a learning agent updates its policy after getting reward feedback the exact ...