The impact of hyperselection on lexicase selection

Thomas Helmuth; Nicholas Freitag McPhee; Lee Spector

doi:10.1145/2908812.2908851

The impact of hyperselection on lexicase selection

Thomas Helmuth, Nicholas Freitag McPhee, Lee Spector

Computer Science (Morris)

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

26 Scopus citations

Abstract

Lexicase selection is a parent selection method that has been shown to improve the problem solving power of genetic programming over a range of problems. Previous work has shown that it can also produce hyperselection events, in which a single individual is selected many more times than other individuals. Here we investigate the role that hyperselection plays in the problem-solving performance of lexicase selection. We run genetic programming on a set of program synthesis benchmark problems using lexicase and tournament selection, confirming that hyperselection occurs significantly more often and more drastically with lexicase selection, which also performs significantly better. We then show results from an experiment indicating that hyperselection is not integral to the problem-solving performance or diversity maintenance observed when using lexicase selection. We conclude that the power of lexicase selection stems from the collection of individuals that it selects, not from the unusual frequencies with which it sometimes selects them.

Original language	English (US)
Title of host publication	GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference
Editors	Tobias Friedrich
Publisher	Association for Computing Machinery, Inc
Pages	717-724
Number of pages	8
ISBN (Electronic)	9781450342063
DOIs	https://doi.org/10.1145/2908812.2908851
State	Published - Jul 20 2016
Event	2016 Genetic and Evolutionary Computation Conference, GECCO 2016 - Denver, United States Duration: Jul 20 2016 → Jul 24 2016

Publication series

Name	GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference

Other

Other	2016 Genetic and Evolutionary Computation Conference, GECCO 2016
Country/Territory	United States
City	Denver
Period	7/20/16 → 7/24/16

Bibliographical note

Funding Information:
Thanks to the members of the Hampshire College Computational Intelligence Lab for discussions that helped shape this work and to Josiah Erikson for systems support. This material is based upon work supported by the National Science Foundation under Grants No. 1129139 and 1331283. Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the National Science Foundation.

Keywords

Hyperselection
Lexicase selection
Program synthesis
Tournament selection

Publisher link

10.1145/2908812.2908851

Find It

Find It at U of M Libraries

Cite this

Helmuth, T., McPhee, N. F., & Spector, L. (2016). The impact of hyperselection on lexicase selection. In T. Friedrich (Ed.), GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference (pp. 717-724). (GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference). Association for Computing Machinery, Inc. https://doi.org/10.1145/2908812.2908851

The impact of hyperselection on lexicase selection. / Helmuth, Thomas; McPhee, Nicholas Freitag; Spector, Lee.
GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference. ed. / Tobias Friedrich. Association for Computing Machinery, Inc, 2016. p. 717-724 (GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference).

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

Helmuth, T, McPhee, NF & Spector, L 2016, The impact of hyperselection on lexicase selection. in T Friedrich (ed.), GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference. GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference, Association for Computing Machinery, Inc, pp. 717-724, 2016 Genetic and Evolutionary Computation Conference, GECCO 2016, Denver, United States, 7/20/16. https://doi.org/10.1145/2908812.2908851

Helmuth T, McPhee NF, Spector L. The impact of hyperselection on lexicase selection. In Friedrich T, editor, GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference. Association for Computing Machinery, Inc. 2016. p. 717-724. (GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference). doi: 10.1145/2908812.2908851

Helmuth, Thomas ; McPhee, Nicholas Freitag ; Spector, Lee. / The impact of hyperselection on lexicase selection. GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference. editor / Tobias Friedrich. Association for Computing Machinery, Inc, 2016. pp. 717-724 (GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference).

@inproceedings{b48e2a4589e24170af5ee17e980f7129,

title = "The impact of hyperselection on lexicase selection",

abstract = "Lexicase selection is a parent selection method that has been shown to improve the problem solving power of genetic programming over a range of problems. Previous work has shown that it can also produce hyperselection events, in which a single individual is selected many more times than other individuals. Here we investigate the role that hyperselection plays in the problem-solving performance of lexicase selection. We run genetic programming on a set of program synthesis benchmark problems using lexicase and tournament selection, confirming that hyperselection occurs significantly more often and more drastically with lexicase selection, which also performs significantly better. We then show results from an experiment indicating that hyperselection is not integral to the problem-solving performance or diversity maintenance observed when using lexicase selection. We conclude that the power of lexicase selection stems from the collection of individuals that it selects, not from the unusual frequencies with which it sometimes selects them.",

keywords = "Hyperselection, Lexicase selection, Program synthesis, Tournament selection",

author = "Thomas Helmuth and McPhee, {Nicholas Freitag} and Lee Spector",

note = "Funding Information: Thanks to the members of the Hampshire College Computational Intelligence Lab for discussions that helped shape this work and to Josiah Erikson for systems support. This material is based upon work supported by the National Science Foundation under Grants No. 1129139 and 1331283. Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the National Science Foundation.; 2016 Genetic and Evolutionary Computation Conference, GECCO 2016 ; Conference date: 20-07-2016 Through 24-07-2016",

year = "2016",

month = jul,

day = "20",

doi = "10.1145/2908812.2908851",

language = "English (US)",

series = "GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference",

publisher = "Association for Computing Machinery, Inc",

pages = "717--724",

editor = "Tobias Friedrich",

booktitle = "GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference",

}

TY - GEN

T1 - The impact of hyperselection on lexicase selection

AU - Helmuth, Thomas

AU - McPhee, Nicholas Freitag

AU - Spector, Lee

N1 - Funding Information: Thanks to the members of the Hampshire College Computational Intelligence Lab for discussions that helped shape this work and to Josiah Erikson for systems support. This material is based upon work supported by the National Science Foundation under Grants No. 1129139 and 1331283. Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the National Science Foundation.

PY - 2016/7/20

Y1 - 2016/7/20

N2 - Lexicase selection is a parent selection method that has been shown to improve the problem solving power of genetic programming over a range of problems. Previous work has shown that it can also produce hyperselection events, in which a single individual is selected many more times than other individuals. Here we investigate the role that hyperselection plays in the problem-solving performance of lexicase selection. We run genetic programming on a set of program synthesis benchmark problems using lexicase and tournament selection, confirming that hyperselection occurs significantly more often and more drastically with lexicase selection, which also performs significantly better. We then show results from an experiment indicating that hyperselection is not integral to the problem-solving performance or diversity maintenance observed when using lexicase selection. We conclude that the power of lexicase selection stems from the collection of individuals that it selects, not from the unusual frequencies with which it sometimes selects them.

AB - Lexicase selection is a parent selection method that has been shown to improve the problem solving power of genetic programming over a range of problems. Previous work has shown that it can also produce hyperselection events, in which a single individual is selected many more times than other individuals. Here we investigate the role that hyperselection plays in the problem-solving performance of lexicase selection. We run genetic programming on a set of program synthesis benchmark problems using lexicase and tournament selection, confirming that hyperselection occurs significantly more often and more drastically with lexicase selection, which also performs significantly better. We then show results from an experiment indicating that hyperselection is not integral to the problem-solving performance or diversity maintenance observed when using lexicase selection. We conclude that the power of lexicase selection stems from the collection of individuals that it selects, not from the unusual frequencies with which it sometimes selects them.

KW - Hyperselection

KW - Lexicase selection

KW - Program synthesis

KW - Tournament selection

UR - http://www.scopus.com/inward/record.url?scp=84985919810&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84985919810&partnerID=8YFLogxK

U2 - 10.1145/2908812.2908851

DO - 10.1145/2908812.2908851

M3 - Conference contribution

AN - SCOPUS:84985919810

T3 - GECCO 2016 - Proceedings of the 2016 Genetic and Evolutionary Computation Conference

SP - 717

EP - 724

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

A2 - Friedrich, Tobias

PB - Association for Computing Machinery, Inc

T2 - 2016 Genetic and Evolutionary Computation Conference, GECCO 2016

Y2 - 20 July 2016 through 24 July 2016

ER -

The impact of hyperselection on lexicase selection

Abstract

Publication series

Other

Bibliographical note

Keywords

Publisher link

Find It

Other files and links

Fingerprint

Cite this