Playing Atari with Deep Reinforcement Learning
This work addresses the challenge of enabling AI to learn from raw sensory data like pixels, which is foundational for advancing autonomous systems in gaming and robotics.
The authors tackled the problem of learning control policies directly from high-dimensional sensory input using reinforcement learning, achieving results that outperformed all previous approaches on six out of seven Atari games and surpassed a human expert on three of them.
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. We apply our method to seven Atari 2600 games from the Arcade Learning Environment, with no adjustment of the architecture or learning algorithm. We find that it outperforms all previous approaches on six of the games and surpasses a human expert on three of them.