YFarmX

On-policy learning

AI

On-policy learning: Training only on data generated by the current policy, which is stable but throws away every earlier sample.

Related terms

Browse the full glossary →