AI Glossary
Reinforcement Learning
AI learning by trial-and-error with reward signals
Definition
Reinforcement learning (RL) is a machine learning paradigm where an AI agent learns by interacting with an environment and receiving reward or penalty signals based on its actions. It has powered breakthroughs in game-playing (AlphaGo, OpenAI Five) and robotics. In the context of LLMs, RL is used in RLHF to align models with human preferences.