Studies in Computational Intelligence

Adaptive Representations for Reinforcement Learning

Authors: Whiteson, Simon

Free Preview
  • Covers recent research in adaptive representations for reinforcement learning
  • Written by leading experts in this field
  • Provides readers with the state of the art on this topic
see more benefits

Buy this book

eBook 95,19 €
price for Spain (gross)
  • ISBN 978-3-642-13932-1
  • Digitally watermarked, DRM-free
  • Included format: PDF
  • ebooks can be used on all reading devices
  • Immediate eBook download after purchase
Hardcover 135,19 €
price for Spain (gross)
  • ISBN 978-3-642-13931-4
  • Free shipping for individuals worldwide
  • Usually dispatched within 3 to 5 business days.
  • The final prices may differ from the prices shown due to specifics of VAT rules
Softcover 116,63 €
price for Spain (gross)
  • ISBN 978-3-642-42231-7
  • Free shipping for individuals worldwide
  • Usually dispatched within 3 to 5 business days.
  • The final prices may differ from the prices shown due to specifics of VAT rules
About this book

This book presents new algorithms for reinforcement learning, a form of machine learning in which an autonomous agent seeks a control policy for a sequential decision task. Since current methods typically rely on manually designed solution representations, agents that automatically adapt their own representations have the potential to dramatically improve performance. This book introduces two novel approaches for automatically discovering high-performing representations. The first approach synthesizes temporal difference methods, the traditional approach to reinforcement learning, with evolutionary methods, which can learn representations for a broad class of optimization problems. This synthesis is accomplished by customizing evolutionary methods to the on-line nature of reinforcement learning and using them to evolve representations for value function approximators. The second approach automatically learns representations based on piecewise-constant approximations of value functions. It begins with coarse representations and gradually refines them during learning, analyzing the current policy and value function to deduce the best refinements. This book also introduces a novel method for devising input representations. This method addresses the feature selection problem by extending an algorithm that evolves the topology and weights of neural networks such that it evolves their inputs too. In addition to introducing these new methods, this book presents extensive empirical results in multiple domains demonstrating that these techniques can substantially improve performance over methods with manual representations.

Table of contents (9 chapters)

  • Introduction

    Whiteson, Shimon

    Pages 1-5

    Preview Buy Chapter 30,19 €
  • Reinforcement Learning

    Whiteson, Shimon

    Pages 7-15

    Preview Buy Chapter 30,19 €
  • On-Line Evolutionary Computation

    Whiteson, Shimon

    Pages 17-30

    Preview Buy Chapter 30,19 €
  • Evolutionary Function Approximation

    Whiteson, Shimon

    Pages 31-46

    Preview Buy Chapter 30,19 €
  • Sample-Efficient Evolutionary Function Approximation

    Whiteson, Shimon

    Pages 47-52

    Preview Buy Chapter 30,19 €

Buy this book

eBook 95,19 €
price for Spain (gross)
  • ISBN 978-3-642-13932-1
  • Digitally watermarked, DRM-free
  • Included format: PDF
  • ebooks can be used on all reading devices
  • Immediate eBook download after purchase
Hardcover 135,19 €
price for Spain (gross)
  • ISBN 978-3-642-13931-4
  • Free shipping for individuals worldwide
  • Usually dispatched within 3 to 5 business days.
  • The final prices may differ from the prices shown due to specifics of VAT rules
Softcover 116,63 €
price for Spain (gross)
  • ISBN 978-3-642-42231-7
  • Free shipping for individuals worldwide
  • Usually dispatched within 3 to 5 business days.
  • The final prices may differ from the prices shown due to specifics of VAT rules
Loading...

Recommended for you

Loading...

Bibliographic Information

Bibliographic Information
Book Title
Adaptive Representations for Reinforcement Learning
Authors
Series Title
Studies in Computational Intelligence
Series Volume
291
Copyright
2010
Publisher
Springer-Verlag Berlin Heidelberg
Copyright Holder
Springer Berlin Heidelberg
eBook ISBN
978-3-642-13932-1
DOI
10.1007/978-3-642-13932-1
Hardcover ISBN
978-3-642-13931-4
Softcover ISBN
978-3-642-42231-7
Series ISSN
1860-949X
Edition Number
1
Number of Pages
XIII, 116
Topics