Reinforcement Learning (RL)

A machine learning paradigm where an agent learns by interacting with an environment and receiving rewards or penalties. RL is used to train game-playing AI, robotics controllers, and to align language models with human preferences via RLHF.