I’m prob. missing sth, but where is the RL part? Isn’t it just a standard supervised learning?