One doc tagged with "Reinforcement Learning"

The paradigm with no dataset — an agent learns by acting, receiving rewards and penalties, and the exploration-exploitation tradeoff that decides whether it finds the best policy or settles for a mediocre one.