Research paper

Q-learning

At a glance

A foundational method for learning which actions lead to the best long-term rewards.

In one paragraph

Watkins and Dayan showed when Q-learning can learn the best actions through trial and error, without first building a model of its environment. The paper proved convergence under stated conditions, formalizing the method introduced in Watkins’s 1989 doctoral work.

Authors 2

Chris Watkins
Peter Dayan

At a glance

In one paragraph

Authors 2

Related works

Playing Atari with Deep Reinforcement Learning→

Human-level control through deep reinforcement learning→

Deep Reinforcement Learning with Double Q-Learning→

Playing Atari with Deep Reinforcement Learning

Human-level control through deep reinforcement learning

Deep Reinforcement Learning with Double Q-Learning