Combining Geometric Semantic GP with Gradient-Descent Optimization

Gloria Pietropolli, Luca Manzoni, Alessia Paoletti, Mauro Castelli

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Citation (Scopus)

Abstract

Geometric semantic genetic programming (GSGP) is a well-known variant of genetic programming (GP) where recombination and mutation operators have a clear semantic effect. Both kind of operators have randomly selected parameters that are not optimized by the search process. In this paper we combine GSGP with a well-known gradient-based optimizer, Adam, in order to leverage the ability of GP to operate structural changes of the individuals with the ability of gradient-based methods to optimize the parameters of a given structure.
Two methods, named HYB-GSGP and HeH-GSGP, are defined and compared with GSGP on a large set of regression problems, showing that the use of Adam can improve the performance on the test set. The idea of merging evolutionary computation and gradient-based optimization is a promising way of combining two methods with very different – and complementary – strengths.
Original languageEnglish
Title of host publicationGenetic Programming. EuroGP 2022
Subtitle of host publication25th European Conference, EuroGP 2022 Held as Part of EvoStar 2022 Madrid, Spain, April 20–22, 2022 Proceedings
EditorsEric Medvet, Gisele Pappa, Bing Xue
Place of PublicationCham
PublisherSpringer
Chapter2
Pages19-33
Number of pages15
ISBN (Electronic)978-3-031-02056-8
ISBN (Print) 978-3-031-02055-1
DOIs
Publication statusPublished - 13 Apr 2022
Event 25th European Conference on Genetic Programming - Virtual
Duration: 20 Apr 202222 Apr 2022
Conference number: 25
http://www.evostar.org/2022/eurogp/

Publication series

NameLecture Notes in Computer Science
Volume13223
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference 25th European Conference on Genetic Programming
Abbreviated titleEuroGP 2022
Period20/04/2222/04/22
Internet address

Fingerprint

Dive into the research topics of 'Combining Geometric Semantic GP with Gradient-Descent Optimization'. Together they form a unique fingerprint.

Cite this