Reinforcement learning is a machine learning approach where an AI agent learns to make decisions by interacting with an environment and receiving feedback in the form of rewards or penalties, rather than learning from a fixed set of labelled examples.

The Core Loop

At each step, the agent observes the current state of its environment, chooses an action, and then receives a reward signal along with a new state. Over many repeated attempts, the agent gradually learns which actions tend to lead to higher rewards in which situations, refining its strategy, often called a policy, through this ongoing cycle of action and feedback.

Where Reinforcement Learning Shines

Reinforcement learning tends to require far more trial-and-error interaction than supervised learning, which makes it well suited to problems where a simulator or repeatable environment is available, but harder to apply directly to situations where mistakes are costly or environments cannot be easily simulated.