Actor-Critic Methods

Twin Delayed Deep Deterministic Policy Gradient

Improvement over DDPG using twin critics to reduce value overestimation and delayed updates of the actor and targets for better stability.

← Zurück