Recently, RL has been gaining more and more popularity. Notably, many of its breakthroughs have come from improvements from supervised methods such as deep learning.
At the moment, most RL algorithms are used in virtual environments such as video games. Luckily, there are companies, such as OpenAI, that have created and released learning environments where it's easy to test the algorithm in different environments.
It's possible to download this learning environment, called Gym, from OpenAI's website.
Additionally, there are real-world applications on RL, and some of them are incredibly impactful. For example, DeepMind, after being used to optimize Google's data centers, was able to reduce the energy consumption and overall energy bill of Google's data centers by 10% and 40%.
A major problem in these algorithms is generalizing the learning...
