Summary

In this chapter, we have explored one of the most important machine learning techniques, RL. We understood the difference between RL and supervised learning. Learning based on behavioral reinforcement for the agent is extremely critical in modeling the intelligent machines that will bridge the gap between human capabilities and the intelligent machines. We have seen the basic concepts of the RL algorithm along with the participating components. We have also tried to establish mathematical equations for a generic RL algorithm where the overall goal is to maximize cumulative rewards for the agent as it transitions through various states with every action.

We have briefly tried to understand the MDPs in a deterministic and stochastic environment. We also explored dynamic programming concepts in brief along with Q-learning and SARSA learning algorithms. In the end, we briefly discussed deep reinforcement learning DRL as a combination of deep neural networks and the reinforcement learning...