Adaptive Structural Hyper-Parameter Configuration by Q-Learning

Haotian Zhang; Jianyong Sun; Zongben Xu

doi:10.1109/CEC48606.2020.9185665

Adaptive Structural Hyper-Parameter Configuration by Q-Learning

Haotian Zhang
, Jianyong Sun
, Zongben Xu

School of Mathematics and Statistics

Xi'an Jiaotong University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

14 Scopus citations

Abstract

Tuning hyper-parameters for evolutionary algorithms is an important issue in computational intelligence. Performance of an evolutionary algorithm depends not only on its operation strategy design, but also on its hyper-parameters. Hyper-parameters can be categorized in two dimensions as structural/numerical and time-invariant/time-variant. Particularly, structural hyper-parameters in existing studies are usually tuned in advance for time-invariant parameters, or with hand-crafted scheduling for time-invariant parameters. In this paper, we make the first attempt to model the tuning of structural hyper-parameters as a reinforcement learning problem, and present to tune the structural hyper-parameter which controls computational resource allocation in the CEC 2018 winner algorithm by Q-learning. Experimental results show favorably against the winner algorithm on the CEC 2018 test functions.

Original language	English
Title of host publication	2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9781728169293
DOIs	https://doi.org/10.1109/CEC48606.2020.9185665
State	Published - Jul 2020
Event	2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Virtual, Glasgow, United Kingdom Duration: 19 Jul 2020 → 24 Jul 2020

Publication series

Name	2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings

Conference

Conference	2020 IEEE Congress on Evolutionary Computation, CEC 2020
Country/Territory	United Kingdom
City	Virtual, Glasgow
Period	19/07/20 → 24/07/20

Keywords

evolutionary algorithm
hyper-parameter tuning
Q-learning
Reinforcement learning

Access to Document

10.1109/CEC48606.2020.9185665

Cite this

Zhang, H., Sun, J., & Xu, Z. (2020). Adaptive Structural Hyper-Parameter Configuration by Q-Learning. In 2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings Article 9185665 (2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CEC48606.2020.9185665

@inproceedings{edca4c0906bc42a7a243eb5bb1f90f1b,

title = "Adaptive Structural Hyper-Parameter Configuration by Q-Learning",

abstract = "Tuning hyper-parameters for evolutionary algorithms is an important issue in computational intelligence. Performance of an evolutionary algorithm depends not only on its operation strategy design, but also on its hyper-parameters. Hyper-parameters can be categorized in two dimensions as structural/numerical and time-invariant/time-variant. Particularly, structural hyper-parameters in existing studies are usually tuned in advance for time-invariant parameters, or with hand-crafted scheduling for time-invariant parameters. In this paper, we make the first attempt to model the tuning of structural hyper-parameters as a reinforcement learning problem, and present to tune the structural hyper-parameter which controls computational resource allocation in the CEC 2018 winner algorithm by Q-learning. Experimental results show favorably against the winner algorithm on the CEC 2018 test functions.",

keywords = "evolutionary algorithm, hyper-parameter tuning, Q-learning, Reinforcement learning",

author = "Haotian Zhang and Jianyong Sun and Zongben Xu",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; 2020 IEEE Congress on Evolutionary Computation, CEC 2020 ; Conference date: 19-07-2020 Through 24-07-2020",

year = "2020",

month = jul,

doi = "10.1109/CEC48606.2020.9185665",

language = "英语",

series = "2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings",

}

Zhang, H, Sun, J & Xu, Z 2020, Adaptive Structural Hyper-Parameter Configuration by Q-Learning. in 2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings., 9185665, 2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings, Institute of Electrical and Electronics Engineers Inc., 2020 IEEE Congress on Evolutionary Computation, CEC 2020, Virtual, Glasgow, United Kingdom, 19/07/20. https://doi.org/10.1109/CEC48606.2020.9185665

Adaptive Structural Hyper-Parameter Configuration by Q-Learning. / Zhang, Haotian; Sun, Jianyong ; Xu, Zongben.
2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings. Institute of Electrical and Electronics Engineers Inc., 2020. 9185665 (2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Adaptive Structural Hyper-Parameter Configuration by Q-Learning

AU - Zhang, Haotian

AU - Sun, Jianyong

AU - Xu, Zongben

PY - 2020/7

Y1 - 2020/7

N2 - Tuning hyper-parameters for evolutionary algorithms is an important issue in computational intelligence. Performance of an evolutionary algorithm depends not only on its operation strategy design, but also on its hyper-parameters. Hyper-parameters can be categorized in two dimensions as structural/numerical and time-invariant/time-variant. Particularly, structural hyper-parameters in existing studies are usually tuned in advance for time-invariant parameters, or with hand-crafted scheduling for time-invariant parameters. In this paper, we make the first attempt to model the tuning of structural hyper-parameters as a reinforcement learning problem, and present to tune the structural hyper-parameter which controls computational resource allocation in the CEC 2018 winner algorithm by Q-learning. Experimental results show favorably against the winner algorithm on the CEC 2018 test functions.

AB - Tuning hyper-parameters for evolutionary algorithms is an important issue in computational intelligence. Performance of an evolutionary algorithm depends not only on its operation strategy design, but also on its hyper-parameters. Hyper-parameters can be categorized in two dimensions as structural/numerical and time-invariant/time-variant. Particularly, structural hyper-parameters in existing studies are usually tuned in advance for time-invariant parameters, or with hand-crafted scheduling for time-invariant parameters. In this paper, we make the first attempt to model the tuning of structural hyper-parameters as a reinforcement learning problem, and present to tune the structural hyper-parameter which controls computational resource allocation in the CEC 2018 winner algorithm by Q-learning. Experimental results show favorably against the winner algorithm on the CEC 2018 test functions.

KW - evolutionary algorithm

KW - hyper-parameter tuning

KW - Q-learning

KW - Reinforcement learning

UR - https://www.scopus.com/pages/publications/85092048070

U2 - 10.1109/CEC48606.2020.9185665

DO - 10.1109/CEC48606.2020.9185665

M3 - 会议稿件

AN - SCOPUS:85092048070

T3 - 2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings

BT - 2020 IEEE Congress on Evolutionary Computation, CEC 2020 - Conference Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2020 IEEE Congress on Evolutionary Computation, CEC 2020

Y2 - 19 July 2020 through 24 July 2020

ER -

Adaptive Structural Hyper-Parameter Configuration by Q-Learning

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this