AI & Machine Learning · Foundational

Reinforcement Learning (RL)

Learning in which an agent takes actions in an environment and improves its policy by maximizing cumulative reward.

Definition

In reinforcement learning an agent observes a state, chooses an action, receives a reward and a new state, and learns through trial and error which actions yield the highest long-term return. It is framed formally as a Markov decision process. RL powers game-playing systems, robotics control and techniques such as RLHF for aligning language models.

Don't confuse it with

Related terms