Concepts1

Policy Gradient Theorem

The policy gradient theorem tells us how to push a stochastic policy’s parameters to increase expected return by following the gradient of expected rewards.