Home / Curated / Algorithms / Deep Q-network learning

Deep Q-network learning

Algorithms AIMech:DeepQNetwork DRAFT

Aliases: DQN, Deep Q-network

A deep Q-learning method that trains a neural action-value function using replayed transitions and a separate target network.

Identity Notes & Concept Scope This record denotes the 2015 DQN learning procedure, not any arbitrary neural Q-function, one Atari checkpoint, or later Double DQN/Rainbow variants.

Claims & Grounded Evidence

1 Claim

Mechanism #learning-method

The 2015 procedure samples replay minibatches, regresses action values toward reward-plus-discounted target-network maxima, and periodically copies the learned network into the target network while collecting actions with an epsilon-greedy policy. Terminal transitions use reward-only targets.

supports Source: Human-level control through deep reinforcement learning · Locator: Methods, Model architecture and Training algorithm for deep Q-networks; Algorithm 1, PDF pp. 6–7 (unnumbered Methods pages)

⚠️ Curated Limitations & Negative Results

The cited formulation selects among discrete actions; continuous-action maximization is not supplied by this algorithm.
The Atari results are not a general human-level intelligence or universal-control claim, and no benchmark result is curated here.

Artificial Intelligence Ontology (AIO) Grounding

Mapped AIO Classes

AIO Target Class	Mapping Relation	Mapping Rationale
`AIO:ReinforcementLearning` Reinforcement Learning	related_match	This algorithm implements reinforcement-learning updates. The AIO target is a learning task, not the same method or an asserted algorithm superclass.

Evidence Sources & Primary Literature

1 Source

Source ID	Publication / Document	Type / Access	Reference & Checksum
`original-paper`	Human-level control through deep reinforcement learning 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	primary paper full_text	ID: doi:10.1038/nature14236 sha256:3ae04e292f28…

Curation Provenance

Curator: Codex

Method: ai_assisted

Model: GPT-6

Created: 2026-10-11T03:44:41Z

Updated: 2026-10-11T03:44:41Z

Source File: data/curated/DeepQNetwork.yaml