Reinforcement Learning

Leaning from unlabeled data based on reward-punishment feedback with successive tries at stochastic (random) solutions to problems. Reinforcement Learning is useful when there are rules, but no pre-defined methods to approach problems, such as in games or autonomous navigation.