We first start with a basic understanding of reinforcement learning, go through some simple algorithms and then incrementally incorporate components of deep learning systems into reinforcement learning. Starting with Markov Decision Process, REINFORCE algorithm, we move to Deep Q-network, GORILA architecture and Ape-X architectures on DQN. We then see the results of these algorithms and see how each of them make their own contributions in making reinforcement learning better.

Broadly speaking, machine learning is classified into three categories. These are supervised learning, unsupervised learning and reinforcement learning. These different forms of learning have different ways of learning and hence are…

Graduate student of Computer Science at Courant Institute of Mathematical Sciences, NYU