Playing Atari with Deep Reinforcement Learning
Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Alex Graves, Ioannis Antonoglou, Daan Wierstra, and Martin Riedmiller submitted Playing Atari with Deep Reinforcement Learning on December 19, 2013. The paper presented a convolutional neural network trained with a variant of Q-learning to learn control policies directly from raw pixel input and evaluated the approach on seven Atari 2600 games from the Arcade Learning Environment.
Historical classification
This verified record belongs to LINEAiGE's artificial-intelligence history layer beginning in 1956.
Why it matters
A documented deep-reinforcement-learning research milestone combining convolutional neural networks with a variant of Q-learning for control directly from high-dimensional visual input.
People, organizations & technologies
PEOPLE
Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Alex Graves, Ioannis Antonoglou, Daan Wierstra, Martin Riedmiller
TECHNOLOGIES
Deep reinforcement learning, Convolutional neural network, Q-learning, Arcade Learning Environment
Evidence & verification
STATE
verified
CONFIDENCE
high
LAST VERIFIED
2026-09-10T06:07:36+08:00
The original arXiv record establishes December 19, 2013 as the first submission date and directly supports the authorship, convolutional-network architecture, variant-of-Q-learning formulation, raw-pixel input, value-function output, and evaluation on seven Atari 2600 games. The date is attached specifically to the paper-submission milestone. The record does not assert the later 'DQN' naming convention, a Nature-publication date, product/software release, deployment, institutional affiliation, geography, ancestry, downstream influence, causality, or relationships.
Sources
Playing Atari with Deep Reinforcement Learning
arXiv
PUBLISHED
2013-12-19
Trace the chronology
Record identity
LINEAiGE IDplaying-atari-deep-rl-2013