Q-learning

E455376

model-free reinforcement learning method reinforcement learning algorithm temporal-difference learning method

Q-learning is a model-free reinforcement learning algorithm that learns an action-value function to optimize decision-making by estimating the expected cumulative reward for each state-action pair.

Try in SPARQL Jump to: Surface forms Disambiguation Statements Elicitation Referenced by

All labels observed (2)

Label	Occurrences
Q-learning canonical	3
Double Q-learning	1

How this entity was disambiguated

This entity first appeared as the object of triple T4586023 — resolving that mention is where its identity was fixed. The disambiguator weighed these candidate entities and picked the highlighted one (or “None”, minting a new entity). This is how homonymy is resolved: the same surface form can point to different entities.

NED1 Entity disambiguation (via context triple) gpt-5-mini-2025-08-07

Target entity: Q-learning
Context triple: [Double DQN, basedOn, Q-learning]

A. Deep Q-Learning
Deep Q-Learning is a reinforcement learning algorithm that uses deep neural networks to approximate Q-values, enabling agents to learn effective policies directly from high-dimensional inputs like raw images.
B. Atari deep Q-network
The Atari deep Q-network is a pioneering deep reinforcement learning system that learned to play a wide range of Atari 2600 video games directly from raw pixels at human-level or better performance.
C. DDPG
DDPG (Deep Deterministic Policy Gradient) is a model-free, off-policy deep reinforcement learning algorithm designed for continuous action spaces, combining ideas from DQN and actor-critic methods.
D. REINFORCE
REINFORCE is a classic Monte Carlo policy gradient algorithm in reinforcement learning that optimizes stochastic policies by estimating gradients from sampled returns.
E. Double DQN
Double DQN is a reinforcement learning algorithm that improves upon standard Deep Q-Networks by reducing overestimation bias through decoupling action selection from action evaluation.
F. None of above. chosen
G. Unsure - the case is ambiguous/there is not enough information to decide.

NED2 Entity disambiguation (via description) gpt-5-mini-2025-08-07

Target entity: Q-learning
Target entity description: Q-learning is a model-free reinforcement learning algorithm that learns an action-value function to optimize decision-making by estimating the expected cumulative reward for each state-action pair.

A. Deep Q-Learning
Deep Q-Learning is a reinforcement learning algorithm that uses deep neural networks to approximate Q-values, enabling agents to learn effective policies directly from high-dimensional inputs like raw images.
B. Atari deep Q-network
The Atari deep Q-network is a pioneering deep reinforcement learning system that learned to play a wide range of Atari 2600 video games directly from raw pixels at human-level or better performance.
C. DDPG
DDPG (Deep Deterministic Policy Gradient) is a model-free, off-policy deep reinforcement learning algorithm designed for continuous action spaces, combining ideas from DQN and actor-critic methods.
D. REINFORCE
REINFORCE is a classic Monte Carlo policy gradient algorithm in reinforcement learning that optimizes stochastic policies by estimating gradients from sampled returns.
E. Double DQN
Double DQN is a reinforcement learning algorithm that improves upon standard Deep Q-Networks by reducing overestimation bias through decoupling action selection from action evaluation.
F. None of above. chosen

Statements (47)

Predicate	Object
instanceOf	model-free reinforcement learning method ⓘ reinforcement learning algorithm ⓘ temporal-difference learning method ⓘ
assumes	discrete action space in basic form ⓘ
canBeExtendedTo	Deep Q-learning ⓘ
canBeImplementedWith	tabular representation ⓘ
canHandle	stochastic rewards ⓘ stochastic transitions ⓘ
canUseExplorationStrategy	epsilon-greedy policy ⓘ softmax action selection ⓘ
canUseFunctionApproximation	linear function approximator ⓘ neural network ⓘ
convergesUnderConditions	Markov decision process ⓘ decaying learning rate ⓘ sufficient exploration ⓘ
describedInPaper	Q-learning ⓘ
differsFrom	SARSA as on-policy vs off-policy ⓘ
doesNotRequire	model of environment dynamics ⓘ
estimates	expected cumulative reward ⓘ
hasAuthor	Christopher J. C. H. Watkins ⓘ
hasCoAuthor	Peter Dayan ⓘ
hasKeyEquation	Q(s,a) ← Q(s,a) + α[r + γ max_{a'} Q(s',a') − Q(s,a)] ⓘ
isModelFree	true ⓘ
isOffPolicy	true ⓘ
isPartOf	reinforcement learning field ⓘ
isRelatedTo	SARSA ⓘ
isSensitiveTo	exploration schedule ⓘ learning rate choice ⓘ reward scaling ⓘ
isUsedFor	optimal policy learning ⓘ
isUsedIn	autonomous decision-making ⓘ game playing ⓘ resource allocation ⓘ robotics control ⓘ
learns	action-value function ⓘ
operatesOn	state-action pairs ⓘ
policyDerivedBy	greedy action selection over Q-values ⓘ
publicationYear	1992 ⓘ
publishedInJournal	Machine Learning ⓘ
requires	reward signal ⓘ
solves	Markov decision process control problems ⓘ
updatesFrom	sample transitions ⓘ
usesDiscountFactor	gamma ⓘ
usesLearningRateParameter	alpha ⓘ
usesMaxOperatorOver	next-state action values ⓘ
usesUpdateRule	Bellman optimality equation ⓘ linked to: Bellman equation
usesValueFunction	true ⓘ

How these facts were elicited

Referenced by (4)

Full triples — surface form annotated when it differs from this entity's canonical label.

Double DQN → basedOn → Q-learning ⓘ

Double DQN → inspiredBy → Double Q-learning ⓘ

linked to: Q-learning

Hado van Hasselt → influencedBy → Q-learning ⓘ

Universal Value Function Approximators → compatibleWith → Q-learning ⓘ

All labels observed (2)

How this entity was disambiguated Show

Statements (47)

How these facts were elicited Show

Referenced by (4)

How this entity was disambiguated

How these facts were elicited