Policy Gradient Methods

Stochastic Policy

A policy represented by a probability distribution π(a|s) over actions, allowing for intrinsic exploration and is essential for policy gradient methods.

← Wstecz