Human-level control through deep reinforcement learning (DQN)
Volodymyr Mnih and colleagues published Human-level control through deep reinforcement learning online in Nature on February 25, 2015. The paper describes a deep Q-network (DQN) agent that combines deep neural networks with reinforcement learning and reports evaluation on 49 Atari 2600 games using pixels and game scores as inputs.
Historical classification
This verified record belongs to LINEAiGE's artificial-intelligence history layer beginning in 1956.
Why it matters
A documented deep-reinforcement-learning milestone establishing the DQN agent and its reported multi-game Atari evaluation.
People, organizations & technologies
PEOPLE
Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A. Rusu, Joel Veness, Marc G. Bellemare, Alex Graves, Martin Riedmiller, Andreas K. Fidjeland, Georg Ostrovski, Stig Petersen, Charles Beattie, Amir Sadik, Ioannis Antonoglou, Helen King, Dharshan Kumaran, Daan Wierstra, Shane Legg, Demis Hassabis
ORGANIZATIONS
Google DeepMind
TECHNOLOGIES
Deep Q-network, DQN, Deep reinforcement learning
Evidence & verification
STATE
verified
CONFIDENCE
high
LAST VERIFIED
2026-09-08T18:45:14+08:00
Nature verifies the title, full author list, online publication date of February 25 2015, Google DeepMind affiliation, DQN terminology, reinforcement-learning/deep-neural-network method and the reported 49-game Atari evaluation. The day-precision date is attached specifically to the Nature online-publication milestone; PubMed records the issue date as February 26 2015. No separate system-birth date, geography, ancestry, influence, causality or relationships are inferred.
Sources
Human-level control through deep reinforcement learning
Nature
PUBLISHED
2015-02-25
Human-level control through deep reinforcement learning
PubMed / U.S. National Library of Medicine
PUBLISHED
2015-02-26
Trace the chronology
Record identity
LINEAiGE IDdqn-2015