Playing Atari with Deep Reinforcement Learning

19 Dec 2013Volodymyr MnihKoray KavukcuogluDavid SilverAlex GravesIoannis AntonoglouDaan WierstraMartin Riedmiller

We present the first deep learning model to successfully learn control policies directly from high-dimensional sensory input using reinforcement learning. The model is a convolutional neural network, trained with a variant of Q-learning, whose input is raw pixels and whose output is a value function estimating future rewards... (read more)

PDF Abstract


Evaluation results from the paper

Task Dataset Model Metric name Metric value Global rank Compare
Atari Games Atari 2600 Beam Rider DQN best Score 5184 # 19
Atari Games Atari 2600 Breakout DQN best Score 225 # 20
Atari Games Atari 2600 Enduro DQN best Score 661 # 14
Atari Games Atari 2600 Pong DQN best Score 21 # 1
Atari Games Atari 2600 Q*Bert DQN best Score 4500 # 21
Atari Games Atari 2600 Seaquest DQN best Score 1740 # 17
Atari Games Atari 2600 Space Invaders DQN best Score 1075 # 20