Сравнение методов

	Q-обучение	Глубокое обучение с подкреплением	Методы градиента политики
Область	Машинное обучение	Глубокое обучение	Машинное обучение
Семейство	Machine learning	Machine learning	Machine learning
Год появления	1992	2015	1992
Автор метода	Christopher Watkins & Peter Dayan	Mnih, V. et al. (DQN)	Ronald Williams (REINFORCE); Sutton et al. (policy gradient theorem)
Тип	Model-free reinforcement-learning control algorithm	Sequential decision-making (agent–environment interaction)	Policy-based reinforcement learning
Основополагающий источник	Watkins, C. J. C. H., & Dayan, P. (1992). Q-learning. Machine Learning, 8(3–4), 279–292.	Mnih, V. et al. (2015). Human-Level Control through Deep Reinforcement Learning. Nature, 518, 529–533.	Williams, R. J. (1992). Simple statistical gradient-following algorithms for connectionist reinforcement learning. Machine Learning, 8(3–4), 229–256.
Другие названия	Q-learning algorithm, tabular Q-learning, off-policy TD control, Q-öğrenme	Derin Pekiştirmeli Öğrenme (DQN / PPO / A3C), derin pekiştirmeli öğrenme, deep RL, DRL	REINFORCE, actor-critic, policy optimization, politika gradyanı
Связанные	3	4	4
Сводка	Q-learning, introduced by Christopher Watkins and Peter Dayan in 1992, is a model-free reinforcement-learning algorithm that learns the value of taking each action in each state — the Q-function — purely from experience, without a model of the environment. It is off-policy: it learns the optimal action-values while following an exploratory behaviour policy, and under standard conditions it provably converges to the optimal policy.	Deep Reinforcement Learning combines neural networks with reinforcement learning so an agent learns by interacting with an environment, popularised by Mnih and colleagues' 2015 Nature work on human-level Atari control. Instead of learning from a fixed labelled dataset, the agent takes actions, observes rewards, and gradually shapes a policy that maximises long-run return.	Policy gradient methods are reinforcement-learning algorithms that optimize a parameterized policy directly by gradient ascent on the expected return, rather than learning action-values and acting greedily. Founded on Ronald Williams' 1992 REINFORCE algorithm and the policy gradient theorem of Sutton and colleagues (2000), they naturally handle stochastic and continuous action spaces and underpin modern actor-critic and deep-RL algorithms.