A computational theory of adaptive behavior based on an evolutionary reinforcement mechanism

J. J. McDowell, Paul L. Soto, Jesse Dallery, Saule Kulubekova

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Two mathematical and two computational theories from the field of human and animal learning are combined to produce a more general theory of adaptive behavior. The cornerstone of this theory is an evolutionary algorithm for reinforcement learning that instantiates the idea that behavior evolves in response to selection pressure from the environment in the form of reinforcement. The evolutionary reinforcement algorithm, along with its associated equilibrium theory, are combined with a mathematical theory of conditioned reinforcement and a computational theory of associative learning that together solve the problem of credit assignment in a biologically plausible way. The result is a biologically-inspired computational theory that enables an artificial organism to adapt continuously to changing environmental conditions and to generate adaptive state-action sequences. Track: Artificial Life, Evolutionary Robotics, Adaptive Behavior

Original languageEnglish (US)
Title of host publicationGECCO 2006 - Genetic and Evolutionary Computation Conference
Pages175-182
Number of pages8
Volume1
StatePublished - 2006
Externally publishedYes
Event8th Annual Genetic and Evolutionary Computation Conference 2006 - Seattle, WA, United States
Duration: Jul 8 2006Jul 12 2006

Other

Other8th Annual Genetic and Evolutionary Computation Conference 2006
CountryUnited States
CitySeattle, WA
Period7/8/067/12/06

Fingerprint

Reinforcement
Reinforcement learning
Evolutionary algorithms
Animals
Robotics

Keywords

  • Adaptive agents
  • Adaptive behavior
  • Conditioned reinforcement
  • Credit assignment
  • Delay-reduction theory
  • Evolutionary algorithms
  • Matching theory
  • Reinforcement learning
  • Rescorla-Wagner rule
  • Stimulus control

ASJC Scopus subject areas

  • Engineering(all)

Cite this

McDowell, J. J., Soto, P. L., Dallery, J., & Kulubekova, S. (2006). A computational theory of adaptive behavior based on an evolutionary reinforcement mechanism. In GECCO 2006 - Genetic and Evolutionary Computation Conference (Vol. 1, pp. 175-182)

A computational theory of adaptive behavior based on an evolutionary reinforcement mechanism. / McDowell, J. J.; Soto, Paul L.; Dallery, Jesse; Kulubekova, Saule.

GECCO 2006 - Genetic and Evolutionary Computation Conference. Vol. 1 2006. p. 175-182.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

McDowell, JJ, Soto, PL, Dallery, J & Kulubekova, S 2006, A computational theory of adaptive behavior based on an evolutionary reinforcement mechanism. in GECCO 2006 - Genetic and Evolutionary Computation Conference. vol. 1, pp. 175-182, 8th Annual Genetic and Evolutionary Computation Conference 2006, Seattle, WA, United States, 7/8/06.
McDowell JJ, Soto PL, Dallery J, Kulubekova S. A computational theory of adaptive behavior based on an evolutionary reinforcement mechanism. In GECCO 2006 - Genetic and Evolutionary Computation Conference. Vol. 1. 2006. p. 175-182
McDowell, J. J. ; Soto, Paul L. ; Dallery, Jesse ; Kulubekova, Saule. / A computational theory of adaptive behavior based on an evolutionary reinforcement mechanism. GECCO 2006 - Genetic and Evolutionary Computation Conference. Vol. 1 2006. pp. 175-182
@inproceedings{7ba10362adbf4c1fa33eb8ecdf4b464e,
title = "A computational theory of adaptive behavior based on an evolutionary reinforcement mechanism",
abstract = "Two mathematical and two computational theories from the field of human and animal learning are combined to produce a more general theory of adaptive behavior. The cornerstone of this theory is an evolutionary algorithm for reinforcement learning that instantiates the idea that behavior evolves in response to selection pressure from the environment in the form of reinforcement. The evolutionary reinforcement algorithm, along with its associated equilibrium theory, are combined with a mathematical theory of conditioned reinforcement and a computational theory of associative learning that together solve the problem of credit assignment in a biologically plausible way. The result is a biologically-inspired computational theory that enables an artificial organism to adapt continuously to changing environmental conditions and to generate adaptive state-action sequences. Track: Artificial Life, Evolutionary Robotics, Adaptive Behavior",
keywords = "Adaptive agents, Adaptive behavior, Conditioned reinforcement, Credit assignment, Delay-reduction theory, Evolutionary algorithms, Matching theory, Reinforcement learning, Rescorla-Wagner rule, Stimulus control",
author = "McDowell, {J. J.} and Soto, {Paul L.} and Jesse Dallery and Saule Kulubekova",
year = "2006",
language = "English (US)",
isbn = "1595931864",
volume = "1",
pages = "175--182",
booktitle = "GECCO 2006 - Genetic and Evolutionary Computation Conference",

}

TY - GEN

T1 - A computational theory of adaptive behavior based on an evolutionary reinforcement mechanism

AU - McDowell, J. J.

AU - Soto, Paul L.

AU - Dallery, Jesse

AU - Kulubekova, Saule

PY - 2006

Y1 - 2006

N2 - Two mathematical and two computational theories from the field of human and animal learning are combined to produce a more general theory of adaptive behavior. The cornerstone of this theory is an evolutionary algorithm for reinforcement learning that instantiates the idea that behavior evolves in response to selection pressure from the environment in the form of reinforcement. The evolutionary reinforcement algorithm, along with its associated equilibrium theory, are combined with a mathematical theory of conditioned reinforcement and a computational theory of associative learning that together solve the problem of credit assignment in a biologically plausible way. The result is a biologically-inspired computational theory that enables an artificial organism to adapt continuously to changing environmental conditions and to generate adaptive state-action sequences. Track: Artificial Life, Evolutionary Robotics, Adaptive Behavior

AB - Two mathematical and two computational theories from the field of human and animal learning are combined to produce a more general theory of adaptive behavior. The cornerstone of this theory is an evolutionary algorithm for reinforcement learning that instantiates the idea that behavior evolves in response to selection pressure from the environment in the form of reinforcement. The evolutionary reinforcement algorithm, along with its associated equilibrium theory, are combined with a mathematical theory of conditioned reinforcement and a computational theory of associative learning that together solve the problem of credit assignment in a biologically plausible way. The result is a biologically-inspired computational theory that enables an artificial organism to adapt continuously to changing environmental conditions and to generate adaptive state-action sequences. Track: Artificial Life, Evolutionary Robotics, Adaptive Behavior

KW - Adaptive agents

KW - Adaptive behavior

KW - Conditioned reinforcement

KW - Credit assignment

KW - Delay-reduction theory

KW - Evolutionary algorithms

KW - Matching theory

KW - Reinforcement learning

KW - Rescorla-Wagner rule

KW - Stimulus control

UR - http://www.scopus.com/inward/record.url?scp=33750228991&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33750228991&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:33750228991

SN - 1595931864

SN - 9781595931863

VL - 1

SP - 175

EP - 182

BT - GECCO 2006 - Genetic and Evolutionary Computation Conference

ER -