Policy Gradient Methods

Entropy Regularization

Adding an entropy term to the objective function to encourage exploration by penalizing overly deterministic policies, improving the robustness of learning.

← Indietro