Step Size Control in Evolutionary Algorithms for Neural Architecture Search

Christian Nieber, Douglas Mota Dias, Enrique Naredo Garcia, Conor Ryan

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

Abstract

This work examines how evolutionary Neural Architecture Search (NAS) algorithms can be improved by controlling the step size of the mutation of numerical parameters. The proposed NAS algorithms are based on F-DENSER, a variation of Dynamic Structured Grammatical Evolution (DSGE). Overall, a (1+5) Evolutionary Strategy is used. Two methods of controlling the step size of mutations of numeric values are compared to Random Search and F-DENSER: Decay of the step size over time and adaptive step size for mutations. The search for lightweight, LeNet-like CNN architectures for MNIST classification is used as a benchmark, optimizing for both accuracy and small architectures. An architecture is described by about 30 evolvable parameters. Experiments show that with step size control, convergence is faster, better performing neural architectures are found on average, and with lower variance. The smallest architecture found during the experiments reached an accuracy of 98.8% on MNIST with only 5,450 free parameters, compared to the 62,158 parameters of LeNet-5.

Original languageEnglish
Title of host publicationProceedings of the 16th International Joint Conference on Computational Intelligence, IJCCI 2024
EditorsFrancesco Marcelloni, Kurosh Madani, Niki van Stein, Joaquim Joaquim
PublisherScience and Technology Publications, Lda
Pages288-295
Number of pages8
ISBN (Print)9789897587214
DOIs
Publication statusPublished - 2024
Event16th International Joint Conference on Computational Intelligence, IJCCI 2024 - Porto, Portugal
Duration: 20 Nov 202422 Nov 2024

Publication series

NameInternational Joint Conference on Computational Intelligence
Volume1
ISSN (Electronic)2184-3236

Conference

Conference16th International Joint Conference on Computational Intelligence, IJCCI 2024
Country/TerritoryPortugal
CityPorto
Period20/11/2422/11/24

Keywords

  • Evolutionary Algorithms
  • LeNet
  • Lightweight CNNs
  • MNIST
  • Neural Architecture Search
  • Step Size Control

Fingerprint

Dive into the research topics of 'Step Size Control in Evolutionary Algorithms for Neural Architecture Search'. Together they form a unique fingerprint.

Cite this