Hybrid Parameter Search and Dynamic Model Selection for Mixed-Variable Bayesian Optimization

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Luo, Hengrui, Cho, Younghyun, Demmel, James W., Li, Xiaoye S., Liu, Yang
Format: Preprint
Veröffentlicht: 2022
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866929214946541568
author Luo, Hengrui
Cho, Younghyun
Demmel, James W.
Li, Xiaoye S.
Liu, Yang
author_facet Luo, Hengrui
Cho, Younghyun
Demmel, James W.
Li, Xiaoye S.
Liu, Yang
contents This paper presents a new type of hybrid model for Bayesian optimization (BO) adept at managing mixed variables, encompassing both quantitative (continuous and integer) and qualitative (categorical) types. Our proposed new hybrid models (named hybridM) merge the Monte Carlo Tree Search structure (MCTS) for categorical variables with Gaussian Processes (GP) for continuous ones. hybridM leverages the upper confidence bound tree search (UCTS) for MCTS strategy, showcasing the tree architecture's integration into Bayesian optimization. Our innovations, including dynamic online kernel selection in the surrogate modeling phase and a unique UCTS search strategy, position our hybrid models as an advancement in mixed-variable surrogate models. Numerical experiments underscore the superiority of hybrid models, highlighting their potential in Bayesian optimization.
format Preprint
id arxiv_https___arxiv_org_abs_2206_01409
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Hybrid Parameter Search and Dynamic Model Selection for Mixed-Variable Bayesian Optimization
Luo, Hengrui
Cho, Younghyun
Demmel, James W.
Li, Xiaoye S.
Liu, Yang
Machine Learning
Statistics Theory
60G15, 62F15, 65C05
This paper presents a new type of hybrid model for Bayesian optimization (BO) adept at managing mixed variables, encompassing both quantitative (continuous and integer) and qualitative (categorical) types. Our proposed new hybrid models (named hybridM) merge the Monte Carlo Tree Search structure (MCTS) for categorical variables with Gaussian Processes (GP) for continuous ones. hybridM leverages the upper confidence bound tree search (UCTS) for MCTS strategy, showcasing the tree architecture's integration into Bayesian optimization. Our innovations, including dynamic online kernel selection in the surrogate modeling phase and a unique UCTS search strategy, position our hybrid models as an advancement in mixed-variable surrogate models. Numerical experiments underscore the superiority of hybrid models, highlighting their potential in Bayesian optimization.
title Hybrid Parameter Search and Dynamic Model Selection for Mixed-Variable Bayesian Optimization
topic Machine Learning
Statistics Theory
60G15, 62F15, 65C05
url https://arxiv.org/abs/2206.01409