Abstract

I present a mathematical, in-depth tutorial about reinforcement learning to the lab members. This was presented to facilitate members to take up RL methods and apply them to their respective problem areas, as well as for myself to understand RL in an in-depth way. The presentation starts off by explaining learning agents from the context of Atari game playing agents,and explains the different cost functions and terminologies used in typical RL methods and papers. This presentation is geared towards enabling the listeners with enough fundamental idea about RL such that they can immediately start reading state-of-art papers about RL and be able to understand the terminology in them.