Faster Convergence with Lexicase Selection in Tree-based Automated Machine Learning
Neural and Evolutionary Computing
2023-02-03 v1 Artificial Intelligence
Abstract
In many evolutionary computation systems, parent selection methods can affect, among other things, convergence to a solution. In this paper, we present a study comparing the role of two commonly used parent selection methods in evolving machine learning pipelines in an automated machine learning system called Tree-based Pipeline Optimization Tool (TPOT). Specifically, we demonstrate, using experiments on multiple datasets, that lexicase selection leads to significantly faster convergence as compared to NSGA-II in TPOT. We also compare the exploration of parts of the search space by these selection methods using a trie data structure that contains information about the pipelines explored in a particular run.
Cite
@article{arxiv.2302.00731,
title = {Faster Convergence with Lexicase Selection in Tree-based Automated Machine Learning},
author = {Nicholas Matsumoto and Anil Kumar Saini and Pedro Ribeiro and Hyunjun Choi and Alena Orlenko and Leo-Pekka Lyytikäinen and Jari O Laurikka and Terho Lehtimäki and Sandra Batista and Jason H. Moore},
journal= {arXiv preprint arXiv:2302.00731},
year = {2023}
}