In a distributed world, consisting of multiple sensors and actuators, there is often a need to automate processes and design strategies to control the "knobs in the world". Examples are:
- How can we configure industrial processes in a factory setting through AI?
- How can we navigate properly in a self-driving car?
- How can we design certain strategies in a competitive game so that we get a better advantage?
Reinforcement learning provides a promising approach to this: based on trying out actions in a clever way, it learns the best strategy for making good decisions. Recently, reinforcement learning has been successfully applied to application domains such as board games (the Google AlphaGo program), robotics, etc.
In this course, we will cover the basics of reinforcement learning and explain the recent breakthroughs in successfully scaling up reinforcement learning to high-dimensional environments (deep reinforcement learning). Moreover, we will explain how reinforcement learning is applied in practical applications and how it can be integrated into distributed environments.