What is Model-based reinforcement learning?

Model-based reinforcement learning

Model-based reinforcement learning: Learning a model of the environment and planning within it, far more sample-efficient than learning purely from real interaction.

Related terms

Inverse reinforcement learning
Model-agnostic meta-learning (MAML)
Offline reinforcement learning
Reinforcement learning (RL)
Reinforcement learning environment (RL environment)
Reinforcement learning from AI feedback (RLAIF)
Reinforcement learning from human feedback (RLHF)
Reinforcement learning with verifiable rewards (RLVR)

Browse the full glossary →