RL4.2 - Basic idea of policy gradient

Basic idea of policy gradient. Policy gradient is a method in Reinforcement Learning and provides an alternative to the well-known Temporal-Difference (TD)-methods. This video present the essential idea of policy gradients.