Reinforcement Learning: Post Quiz

How does RL training algorithm knows how well it did?