Skip to content
Sign in
All terms
Glossary · Term

Reinforcement Learning

L3 · AdvancedFoundational

In plain terms

Learning to act from rewards through trial and error.

Technical definition

A paradigm where an agent chooses actions in an environment to maximize cumulative reward, learning a policy or value function from the outcomes of its own actions.

Example

An agent learning a maze by getting reward only when it reaches the goal.

Try it in the Q-Learning Gridworld labExplore in the AI Universe

Related terms