Advanced Reinforcement Learning

Nathaniel D. Daw

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

This chapter reviews issues of current research in reinforcement learning theories and their neural substrates. We consider how the formal constructs of states, actions, and rewards that these theories describe can be understood to map onto counterparts experienced by biological organisms learning in the real world. In each case, this correspondence involves significant difficulties. However, elaborated theoretical accounts from computer science clarify, in each case, how to extend these theories to more realistic circumstances while still preserving the core prediction error-driven learning mechanism that has been prominent in neuroeconomic accounts.

Original languageEnglish (US)
Title of host publicationNeuroeconomics: Decision Making and the Brain: Second Edition
PublisherElsevier Inc.
Pages299-320
Number of pages22
ISBN (Print)9780124160088
DOIs
StatePublished - Sep 2013

Fingerprint

Learning
Reward
Research
Reinforcement (Psychology)

Keywords

  • Dopamine
  • Hierarchical reinforcement learning
  • Reinforcement learning
  • Uncertainty

ASJC Scopus subject areas

  • Neuroscience(all)

Cite this

Daw, N. D. (2013). Advanced Reinforcement Learning. In Neuroeconomics: Decision Making and the Brain: Second Edition (pp. 299-320). Elsevier Inc.. https://doi.org/10.1016/B978-0-12-416008-8.00016-4

Advanced Reinforcement Learning. / Daw, Nathaniel D.

Neuroeconomics: Decision Making and the Brain: Second Edition. Elsevier Inc., 2013. p. 299-320.

Research output: Chapter in Book/Report/Conference proceedingChapter

Daw, ND 2013, Advanced Reinforcement Learning. in Neuroeconomics: Decision Making and the Brain: Second Edition. Elsevier Inc., pp. 299-320. https://doi.org/10.1016/B978-0-12-416008-8.00016-4
Daw ND. Advanced Reinforcement Learning. In Neuroeconomics: Decision Making and the Brain: Second Edition. Elsevier Inc. 2013. p. 299-320 https://doi.org/10.1016/B978-0-12-416008-8.00016-4
Daw, Nathaniel D. / Advanced Reinforcement Learning. Neuroeconomics: Decision Making and the Brain: Second Edition. Elsevier Inc., 2013. pp. 299-320
@inbook{86cb75dd0cb94e3c805b69a201b3c6a3,
title = "Advanced Reinforcement Learning",
abstract = "This chapter reviews issues of current research in reinforcement learning theories and their neural substrates. We consider how the formal constructs of states, actions, and rewards that these theories describe can be understood to map onto counterparts experienced by biological organisms learning in the real world. In each case, this correspondence involves significant difficulties. However, elaborated theoretical accounts from computer science clarify, in each case, how to extend these theories to more realistic circumstances while still preserving the core prediction error-driven learning mechanism that has been prominent in neuroeconomic accounts.",
keywords = "Dopamine, Hierarchical reinforcement learning, Reinforcement learning, Uncertainty",
author = "Daw, {Nathaniel D.}",
year = "2013",
month = "9",
doi = "10.1016/B978-0-12-416008-8.00016-4",
language = "English (US)",
isbn = "9780124160088",
pages = "299--320",
booktitle = "Neuroeconomics: Decision Making and the Brain: Second Edition",
publisher = "Elsevier Inc.",

}

TY - CHAP

T1 - Advanced Reinforcement Learning

AU - Daw, Nathaniel D.

PY - 2013/9

Y1 - 2013/9

N2 - This chapter reviews issues of current research in reinforcement learning theories and their neural substrates. We consider how the formal constructs of states, actions, and rewards that these theories describe can be understood to map onto counterparts experienced by biological organisms learning in the real world. In each case, this correspondence involves significant difficulties. However, elaborated theoretical accounts from computer science clarify, in each case, how to extend these theories to more realistic circumstances while still preserving the core prediction error-driven learning mechanism that has been prominent in neuroeconomic accounts.

AB - This chapter reviews issues of current research in reinforcement learning theories and their neural substrates. We consider how the formal constructs of states, actions, and rewards that these theories describe can be understood to map onto counterparts experienced by biological organisms learning in the real world. In each case, this correspondence involves significant difficulties. However, elaborated theoretical accounts from computer science clarify, in each case, how to extend these theories to more realistic circumstances while still preserving the core prediction error-driven learning mechanism that has been prominent in neuroeconomic accounts.

KW - Dopamine

KW - Hierarchical reinforcement learning

KW - Reinforcement learning

KW - Uncertainty

UR - http://www.scopus.com/inward/record.url?scp=84897397355&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84897397355&partnerID=8YFLogxK

U2 - 10.1016/B978-0-12-416008-8.00016-4

DO - 10.1016/B978-0-12-416008-8.00016-4

M3 - Chapter

SN - 9780124160088

SP - 299

EP - 320

BT - Neuroeconomics: Decision Making and the Brain: Second Edition

PB - Elsevier Inc.

ER -