Summary
My congratulations, you have made another step towards understanding modern, state-of-the-art RL methods! In this chapter, you learned about some very important concepts that are widely used in deep RL: the value of the state, the value of the action, and the Bellman equation in various forms.
We also covered the value iteration method, which is a very important building block in the area of Q-learning. Finally, you got to know how value iteration can improve our FrozenLake solution.
In the next chapter, you will learn about deep Q-networks, which started the deep RL revolution in 2013 by beating humans on lots of Atari 2600 games.