Curious: Searching for unknown regions of space with a subpopulation-based algorithm

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

Abstract

Intrinsic motivation and novelty search are promising approaches to deal with plateaus, deceptive functions and other exploration problems where using only the main objective function is insufficient. However, it is not clear until now how and if intrinsic motivation (novelty search) can improve single objective algorithms in general. The hurdle is that using multi-objective algorithms to deal with single-objective problems adds an unnecessary overhead such as the search for non-dominated solutions. Here, we propose the Curious algorithm which is the first multi-objective algorithm focused on solving single-objective problems. Curious uses two subpopulations algorithms. One subpopulation is dedicated for improving objective function values and another one is added to search for unknown regions of space based on objective prediction errors. By using a differential evolution operator, genes from individuals in all subpopulations are mixed. In this way, the promising regions (solutions with high fitness) and unknown regions (solutions with high prediction error) are searched simultaneously. Because of thus realized strong yet well controlled novelty search, the algorithm possesses powerful exploration ability and outperforms usual single population based algorithms such as differential evolution. Thus, it demonstrates that the addition of intrinsic motivation is promising and should improve further single objective algorithms in general.

Original languageEnglish
Title of host publicationGECCO 2016 Companion - Proceedings of the 2016 Genetic and Evolutionary Computation Conference
PublisherAssociation for Computing Machinery, Inc
Pages145-146
Number of pages2
ISBN (Electronic)9781450343237
DOIs
Publication statusPublished - Jul 20 2016
Event2016 Genetic and Evolutionary Computation Conference, GECCO 2016 Companion - Denver, United States
Duration: Jul 20 2016Jul 24 2016

Other

Other2016 Genetic and Evolutionary Computation Conference, GECCO 2016 Companion
CountryUnited States
CityDenver
Period7/20/167/24/16

Fingerprint

Mathematical operators
Genes

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Science Applications
  • Computational Theory and Mathematics

Cite this

Vargas, D. V., & Murata, J. (2016). Curious: Searching for unknown regions of space with a subpopulation-based algorithm. In GECCO 2016 Companion - Proceedings of the 2016 Genetic and Evolutionary Computation Conference (pp. 145-146). Association for Computing Machinery, Inc. https://doi.org/10.1145/2908961.2908982

Curious : Searching for unknown regions of space with a subpopulation-based algorithm. / Vargas, Danilo Vasconcellos; Murata, Junichi.

GECCO 2016 Companion - Proceedings of the 2016 Genetic and Evolutionary Computation Conference. Association for Computing Machinery, Inc, 2016. p. 145-146.

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

Vargas, DV & Murata, J 2016, Curious: Searching for unknown regions of space with a subpopulation-based algorithm. in GECCO 2016 Companion - Proceedings of the 2016 Genetic and Evolutionary Computation Conference. Association for Computing Machinery, Inc, pp. 145-146, 2016 Genetic and Evolutionary Computation Conference, GECCO 2016 Companion, Denver, United States, 7/20/16. https://doi.org/10.1145/2908961.2908982
Vargas DV, Murata J. Curious: Searching for unknown regions of space with a subpopulation-based algorithm. In GECCO 2016 Companion - Proceedings of the 2016 Genetic and Evolutionary Computation Conference. Association for Computing Machinery, Inc. 2016. p. 145-146 https://doi.org/10.1145/2908961.2908982
Vargas, Danilo Vasconcellos ; Murata, Junichi. / Curious : Searching for unknown regions of space with a subpopulation-based algorithm. GECCO 2016 Companion - Proceedings of the 2016 Genetic and Evolutionary Computation Conference. Association for Computing Machinery, Inc, 2016. pp. 145-146
@inproceedings{d5e6af8360a346ceb71d8c9a6fe80380,
title = "Curious: Searching for unknown regions of space with a subpopulation-based algorithm",
abstract = "Intrinsic motivation and novelty search are promising approaches to deal with plateaus, deceptive functions and other exploration problems where using only the main objective function is insufficient. However, it is not clear until now how and if intrinsic motivation (novelty search) can improve single objective algorithms in general. The hurdle is that using multi-objective algorithms to deal with single-objective problems adds an unnecessary overhead such as the search for non-dominated solutions. Here, we propose the Curious algorithm which is the first multi-objective algorithm focused on solving single-objective problems. Curious uses two subpopulations algorithms. One subpopulation is dedicated for improving objective function values and another one is added to search for unknown regions of space based on objective prediction errors. By using a differential evolution operator, genes from individuals in all subpopulations are mixed. In this way, the promising regions (solutions with high fitness) and unknown regions (solutions with high prediction error) are searched simultaneously. Because of thus realized strong yet well controlled novelty search, the algorithm possesses powerful exploration ability and outperforms usual single population based algorithms such as differential evolution. Thus, it demonstrates that the addition of intrinsic motivation is promising and should improve further single objective algorithms in general.",
author = "Vargas, {Danilo Vasconcellos} and Junichi Murata",
year = "2016",
month = "7",
day = "20",
doi = "10.1145/2908961.2908982",
language = "English",
pages = "145--146",
booktitle = "GECCO 2016 Companion - Proceedings of the 2016 Genetic and Evolutionary Computation Conference",
publisher = "Association for Computing Machinery, Inc",

}

TY - GEN

T1 - Curious

T2 - Searching for unknown regions of space with a subpopulation-based algorithm

AU - Vargas, Danilo Vasconcellos

AU - Murata, Junichi

PY - 2016/7/20

Y1 - 2016/7/20

N2 - Intrinsic motivation and novelty search are promising approaches to deal with plateaus, deceptive functions and other exploration problems where using only the main objective function is insufficient. However, it is not clear until now how and if intrinsic motivation (novelty search) can improve single objective algorithms in general. The hurdle is that using multi-objective algorithms to deal with single-objective problems adds an unnecessary overhead such as the search for non-dominated solutions. Here, we propose the Curious algorithm which is the first multi-objective algorithm focused on solving single-objective problems. Curious uses two subpopulations algorithms. One subpopulation is dedicated for improving objective function values and another one is added to search for unknown regions of space based on objective prediction errors. By using a differential evolution operator, genes from individuals in all subpopulations are mixed. In this way, the promising regions (solutions with high fitness) and unknown regions (solutions with high prediction error) are searched simultaneously. Because of thus realized strong yet well controlled novelty search, the algorithm possesses powerful exploration ability and outperforms usual single population based algorithms such as differential evolution. Thus, it demonstrates that the addition of intrinsic motivation is promising and should improve further single objective algorithms in general.

AB - Intrinsic motivation and novelty search are promising approaches to deal with plateaus, deceptive functions and other exploration problems where using only the main objective function is insufficient. However, it is not clear until now how and if intrinsic motivation (novelty search) can improve single objective algorithms in general. The hurdle is that using multi-objective algorithms to deal with single-objective problems adds an unnecessary overhead such as the search for non-dominated solutions. Here, we propose the Curious algorithm which is the first multi-objective algorithm focused on solving single-objective problems. Curious uses two subpopulations algorithms. One subpopulation is dedicated for improving objective function values and another one is added to search for unknown regions of space based on objective prediction errors. By using a differential evolution operator, genes from individuals in all subpopulations are mixed. In this way, the promising regions (solutions with high fitness) and unknown regions (solutions with high prediction error) are searched simultaneously. Because of thus realized strong yet well controlled novelty search, the algorithm possesses powerful exploration ability and outperforms usual single population based algorithms such as differential evolution. Thus, it demonstrates that the addition of intrinsic motivation is promising and should improve further single objective algorithms in general.

UR - http://www.scopus.com/inward/record.url?scp=84986264888&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84986264888&partnerID=8YFLogxK

U2 - 10.1145/2908961.2908982

DO - 10.1145/2908961.2908982

M3 - Conference contribution

AN - SCOPUS:84986264888

SP - 145

EP - 146

BT - GECCO 2016 Companion - Proceedings of the 2016 Genetic and Evolutionary Computation Conference

PB - Association for Computing Machinery, Inc

ER -