https://devslem.github.io/reinforcement-learning/rl-fundamental/multi-armed-bandits/