Automated Class Imbalance Learning via Few-Shot Multi-Objective Bayesian Optimization With Deep Kernel Gaussian Processes

Automated Class Imbalance Learning (AutoCIL) is an emerging paradigm that leverages Combined Algorithm Selection and Hyperparameter Optimization (CASH) to automate the configuration of resampling strategies and classifiers for imbalanced classification tasks. Existing AutoCIL methods focus solely on...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhaoyang Wang, Shuo Wang, Damien Ernst, Chenguang Xiao
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11087233/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849421043664420864
author Zhaoyang Wang
Shuo Wang
Damien Ernst
Chenguang Xiao
author_facet Zhaoyang Wang
Shuo Wang
Damien Ernst
Chenguang Xiao
author_sort Zhaoyang Wang
collection DOAJ
description Automated Class Imbalance Learning (AutoCIL) is an emerging paradigm that leverages Combined Algorithm Selection and Hyperparameter Optimization (CASH) to automate the configuration of resampling strategies and classifiers for imbalanced classification tasks. Existing AutoCIL methods focus solely on single-objective optimization. However, real-world applications often involve multiple, conflicting objectives—such as predictive performance and computational cost—that must be jointly optimized. Ignoring such trade-offs limits the adaptability and practicality of current methods. In this work, we propose a novel approach called AutoCIL-FMOBO (AutoCIL via Few-shot Multi-Objective Bayesian Optimization). Specifically, we design meta-learned deep kernel Gaussian process surrogates trained on a meta-dataset constructed from pre-evaluated results obtained by running configurations in the search space on class-imbalanced datasets. Then, these surrogate models with prior optimization knowledge are combined with the Expected Hypervolume Improvement (EHVI) acquisition function in a Bayesian optimization framework to efficiently discover Pareto-optimal configurations for the target task, which enables AutoCIL-FMOBO to jointly optimize key components, such as resampling methods, classifiers, and their hyperparameters, under a multi-objective setting. Experimental results on 15 real-world class-imbalanced datasets demonstrate that our approach outperforms baselines in both effectiveness and sample efficiency, while maintaining generalization across tasks and achieving competitive performance under a multi-objective setting.
format Article
id doaj-art-daef72118ef94fd8a4178cd84760f526
institution Kabale University
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-daef72118ef94fd8a4178cd84760f5262025-08-20T03:31:34ZengIEEEIEEE Access2169-35362025-01-011313183913185510.1109/ACCESS.2025.359103411087233Automated Class Imbalance Learning via Few-Shot Multi-Objective Bayesian Optimization With Deep Kernel Gaussian ProcessesZhaoyang Wang0https://orcid.org/0000-0001-7701-1128Shuo Wang1https://orcid.org/0000-0003-1380-6428Damien Ernst2https://orcid.org/0000-0002-3035-8260Chenguang Xiao3School of Computer Science, University of Birmingham, Birmingham, U.K.School of Computer Science, University of Birmingham, Birmingham, U.K.Department of Electrical Engineering and Computer Science, University of Liège, Liège, BelgiumSchool of Computer Science, University of Birmingham, Birmingham, U.K.Automated Class Imbalance Learning (AutoCIL) is an emerging paradigm that leverages Combined Algorithm Selection and Hyperparameter Optimization (CASH) to automate the configuration of resampling strategies and classifiers for imbalanced classification tasks. Existing AutoCIL methods focus solely on single-objective optimization. However, real-world applications often involve multiple, conflicting objectives—such as predictive performance and computational cost—that must be jointly optimized. Ignoring such trade-offs limits the adaptability and practicality of current methods. In this work, we propose a novel approach called AutoCIL-FMOBO (AutoCIL via Few-shot Multi-Objective Bayesian Optimization). Specifically, we design meta-learned deep kernel Gaussian process surrogates trained on a meta-dataset constructed from pre-evaluated results obtained by running configurations in the search space on class-imbalanced datasets. Then, these surrogate models with prior optimization knowledge are combined with the Expected Hypervolume Improvement (EHVI) acquisition function in a Bayesian optimization framework to efficiently discover Pareto-optimal configurations for the target task, which enables AutoCIL-FMOBO to jointly optimize key components, such as resampling methods, classifiers, and their hyperparameters, under a multi-objective setting. Experimental results on 15 real-world class-imbalanced datasets demonstrate that our approach outperforms baselines in both effectiveness and sample efficiency, while maintaining generalization across tasks and achieving competitive performance under a multi-objective setting.https://ieeexplore.ieee.org/document/11087233/Class imbalance learningmulti-objective optimizationBayesian optimizationautomated machine learningmeta-learningGaussian processes
spellingShingle Zhaoyang Wang
Shuo Wang
Damien Ernst
Chenguang Xiao
Automated Class Imbalance Learning via Few-Shot Multi-Objective Bayesian Optimization With Deep Kernel Gaussian Processes
IEEE Access
Class imbalance learning
multi-objective optimization
Bayesian optimization
automated machine learning
meta-learning
Gaussian processes
title Automated Class Imbalance Learning via Few-Shot Multi-Objective Bayesian Optimization With Deep Kernel Gaussian Processes
title_full Automated Class Imbalance Learning via Few-Shot Multi-Objective Bayesian Optimization With Deep Kernel Gaussian Processes
title_fullStr Automated Class Imbalance Learning via Few-Shot Multi-Objective Bayesian Optimization With Deep Kernel Gaussian Processes
title_full_unstemmed Automated Class Imbalance Learning via Few-Shot Multi-Objective Bayesian Optimization With Deep Kernel Gaussian Processes
title_short Automated Class Imbalance Learning via Few-Shot Multi-Objective Bayesian Optimization With Deep Kernel Gaussian Processes
title_sort automated class imbalance learning via few shot multi objective bayesian optimization with deep kernel gaussian processes
topic Class imbalance learning
multi-objective optimization
Bayesian optimization
automated machine learning
meta-learning
Gaussian processes
url https://ieeexplore.ieee.org/document/11087233/
work_keys_str_mv AT zhaoyangwang automatedclassimbalancelearningviafewshotmultiobjectivebayesianoptimizationwithdeepkernelgaussianprocesses
AT shuowang automatedclassimbalancelearningviafewshotmultiobjectivebayesianoptimizationwithdeepkernelgaussianprocesses
AT damienernst automatedclassimbalancelearningviafewshotmultiobjectivebayesianoptimizationwithdeepkernelgaussianprocesses
AT chenguangxiao automatedclassimbalancelearningviafewshotmultiobjectivebayesianoptimizationwithdeepkernelgaussianprocesses