Reinforcement Learning: An Introduction

Reinforcement learning is a computational approach to learning where an agent tries to maximize reward in a complex environment. Richard Sutton and Andrew Barto's new edition of their widely used text provides a clear and simple account of the field's key ideas and algorithms, with new topics and updated coverage of other topics. Part I covers core, online learning algorithms, while Part II extends these ideas to function approximation and offers expanded treatment of off-policy learning and policy-gradient methods. Part III has new chapters on reinforcement learning's relationships to psychology and neuroscience, as well as an updated case-studies chapter including AlphaGo and AlphaGo Zero, Atari game playing, and IBM Watson's wagering strategy. The final chapter discusses the future societal impacts of reinforcement learning.

Format: Hardback
Length: 552 pages
Publication date: 13 November 2018
Publisher: MIT Press Ltd

Reinforcement learning (RL) is a computational approach to learning where an agent seeks to maximize its total reward while interacting with a complex and uncertain environment. In the second edition of their widely used text on RL, Richard Sutton and Andrew Barto provide a clear and concise explanation of the field's key ideas and algorithms. This edition has been significantly expanded and updated, covering new topics and updating coverage of existing ones. The book focuses on core, online learning algorithms, with the more mathematical material presented in shaded boxes. Part I covers as much of RL as possible without going beyond the tabular case for which exact solutions can be found. Many algorithms presented in this part are new for the second edition, including UCB, Expected Sarsa, and Double Learning. Part II extends these ideas to function approximation, with new sections on artificial neural networks and the Fourier basis, and offers expanded treatment of off-policy learning and policy-gradient methods. Part III has new chapters on RL's relationships to psychology and neuroscience, as well as an updated case-studies chapter including AlphaGo and AlphaGo Zero, Atari game playing, and IBM Watson's wagering strategy. The final chapter discusses the future societal impacts of RL.

Weight: 1208g
Dimension: 187 x 236 x 33 (mm)
ISBN-13: 9780262039246
Edition number: second edition