AI Glossary

Temporal difference learning

Also called TD learning

What does “Temporal difference learning” mean?

Temporal difference learning: Learning predictions using the difference between a current estimate and a later observation plus a new estimate.

Temporal difference learning visual guide: Learning predictions using the difference between a current estimate and a later observation plus a new estimate.
Visual guide Open full size (new tab)

What is an example of “Temporal difference learning”?

An agent updates a state's value after taking a step and seeing the next state.

Keep learning

See also

Reinforcement learningTermReward modelTermReward hackingTermProximal policy optimizationTermQ-learningTerm

Same topic

Reward shapingTermTransfer datasetTermSynthetic supervisionTermTransfer learningTermSynthetic dataTerm