policy gradient algorithms

1. sample actions
2. observe rewards
3. tweak the policy

Sources

https://towardsdatascience.com/policy-gradients-in-reinforcement-learning-explained-ecec7df94245