Linked from

The 21 pages that link to Reinforcement learning, each with the reason it gives.

• Artificial intelligenceBroader topic: It trains agents to choose actions by evaluating their consequences.

• Machine learningBroader topic: It addresses decisions whose consequences unfold through interaction.

• RoboticsRelated: Some robots learn control policies through repeated interaction with environments.

• HabitNarrower topic: Reward-based learning helps explain how repeated actions become associated with contexts.

• Basal gangliaRelated: Dopamine signals help basal ganglia circuits update action choices from outcomes.

• LearningRelated: It formalizes reward-guided learning in artificial agents.

• Social learningRelated: Social cues can supply information about rewards even before direct experience.

• Artificial general intelligenceRelated: It offers methods for learning behavior through interaction and feedback.

• Evolutionary game theoryCompared with: It can produce adaptation without biological reproduction, though its update mechanism differs.

• Game stateNarrower topic: Agents use state descriptions to choose actions and learn from outcomes.

• Optimal controlCompared with: It can approximate optimal policies when dynamics or costs are unknown, but typically learns from data.

• Training dataCompared with: Its learning experience may consist of interaction trajectories rather than a fixed labeled dataset.

• Gambling disorderNarrower topic: It provides a broader account of how gambling outcomes shape repeated behavior.

• Computational neuroscienceRelated: It provides mathematical tools for modeling reward-guided learning in brains.

• Artificial intelligence in video gamesRelated: It can train agents through repeated gameplay rather than hand-authored rules.

• Richard BellmanRelated: Its value functions and learning algorithms draw directly on Bellman’s equations.

• Tic-tac-toeNarrower topic: Tic-tac-toe provides a simple environment for learning move choices through repeated play.

• Andrew NgCompared with: It is a distinct machine-learning approach from the supervised-learning methods central to many of Ng’s courses.

• Computer GoRelated: Self-play lets Go programs learn strategies from the consequences of their moves.

• Demis HassabisRelated: It underlies many DeepMind systems developed during Hassabis’s leadership.

• NeuroeconomicsRelated: Its prediction errors link reward experience to changes in later choices.