Richard S. Sutton

Learn More
The reinforcement learning (RL) problem is the challenge of artificial intelligence in a microcosm; how can we build an agent that can plan, learn, perceive, and act in a complex world? There’s a great new book on the market that lays out the conceptual and algorithmic foundations of this exciting area. RL pioneers Rich Sutton and Andy Barto have published(More)
This article introduces a class of incremental learning procedures specialized for prediction-that is, for using past experience with an incompletely known system to predict its future behavior. Whereas conventional prediction-learning methods assign credit by means of the difference between predicted and actual outcomes, the new methods assign credit by(More)
Function approximation is essential to reinforcement learning, but the standard approach of approximating a value function and determining a policy from it has so far proven theoretically intractable. In this paper we explore an alternative approach in which the policy is explicitly represented by its own function approximator, independent of the value(More)
Learning, planning, and representing knowledge at multiple levels of temporal abstraction are key, longstanding challenges for AI. In this paper we consider how these challenges can be addressed within the mathematical framework of reinforcement learning and Markov decision processes (MDPs). We extend the usual notion of action in this framework to include(More)
Many adaptive neural network theories are based on neuronlike adaptive elements that can behave as single unit analogs of associative conditioning. In this article we develop a similar adaptive element, but one which is more closely in accord with the facts of animal learning theory than elements commonly studied in adaptive network research. We suggest(More)
This paper extends previous work with Dyna a class of architectures for intelligent systems based on approximating dynamic program ming methods Dyna architectures integrate trial and error reinforcement learning and execution time planning into a single process operating alternately on the world and on a learned model of the world In this paper I present(More)
On large problems, reinforcement learning systems must use parameterized function approximators such as neural networks in order to generalize between similar situations and actions. In these cases there are no strong theoretical results on the accuracy of convergence, and computational results have been mixed. In particular, Boyan and Moore reported at(More)
Eligibility traces have been shown to speed reinforcement learning, to make it more robust to hidden states, and to provide a link between Monte Carlo and temporal-difference methods. Here we generalize eligibility traces to off-policy learning, in which one learns about a policy different from the policy that generates the data. Off-policy methods can(More)