How to use reinforcement learning to play tic-tac-toe

A simple tutorial about Q-learning

Q-learning is a brilliant and fundamental method of reinforcement learning that has shown a lot of success recently thanks to the deep learning revolution. Although this tutorial will not explain what is known as deep Q-learning, we will go through the original Q-learning algorithm to teach an agent how to play a game of tic-tac-toe. Despite its simplicity, we will see that it can give excellent results.