Guardado en:
Detalles Bibliográficos
Autores principales: Ip, Joshua Hang Sai, Makrygiorgos, Georgios, Mesbah, Ali
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:https://arxiv.org/abs/2601.21357
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866914577096114176
author Ip, Joshua Hang Sai
Makrygiorgos, Georgios
Mesbah, Ali
author_facet Ip, Joshua Hang Sai
Makrygiorgos, Georgios
Mesbah, Ali
contents Bayesian Optimization (BO) is a principled framework for optimizing expensive black-box functions, with Expected Improvement (EI) among its most widely used acquisition functions. Despite its empirical success, EI is agnostic to first-order optimality conditions, relying solely on objective-value improvement. As a result, it can exhibit vanishing acquisition signals where the improvement criterion is uninformative, limiting its effectiveness in guiding search. We propose Expected Improvement via Gradient Norms (EI-GN), a novel acquisition function that extends the improvement principle to incorporate first-order stationarity, promoting sampling in regions that are both high-performing and close to stationary points. We derive a tractable closed-form expression for EI-GN and show that it remains consistent with the improvement-based acquisition framework. By embedding progress toward stationarity into the acquisition criterion, EI-GN provides a richer and more informative notion of improvement. Empirical results on standard BO benchmarks demonstrate consistent gains over baseline methods, and we further illustrate its applicability to control policy learning.
format Preprint
id arxiv_https___arxiv_org_abs_2601_21357
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Beyond Objective-Based Improvement: Stationarity-Aware Expected Improvement for Bayesian Optimization
Ip, Joshua Hang Sai
Makrygiorgos, Georgios
Mesbah, Ali
Machine Learning
Bayesian Optimization (BO) is a principled framework for optimizing expensive black-box functions, with Expected Improvement (EI) among its most widely used acquisition functions. Despite its empirical success, EI is agnostic to first-order optimality conditions, relying solely on objective-value improvement. As a result, it can exhibit vanishing acquisition signals where the improvement criterion is uninformative, limiting its effectiveness in guiding search. We propose Expected Improvement via Gradient Norms (EI-GN), a novel acquisition function that extends the improvement principle to incorporate first-order stationarity, promoting sampling in regions that are both high-performing and close to stationary points. We derive a tractable closed-form expression for EI-GN and show that it remains consistent with the improvement-based acquisition framework. By embedding progress toward stationarity into the acquisition criterion, EI-GN provides a richer and more informative notion of improvement. Empirical results on standard BO benchmarks demonstrate consistent gains over baseline methods, and we further illustrate its applicability to control policy learning.
title Beyond Objective-Based Improvement: Stationarity-Aware Expected Improvement for Bayesian Optimization
topic Machine Learning
url https://arxiv.org/abs/2601.21357