Jack What is an action in reinforcement learning, and how does it represent a decision made by an agent? How does choosing different actions based on the current state affect rewards and help the agent learn an effective strategy?