Multi-Objective Bayesian Optimization via Adaptive ε \varepsilon -Constraints Decomposition

Yaohong Yang, Sammie Katt, Samuel Kaski · 2026

Read Multi-Objective Bayesian Optimization via Adaptive ε \varepsilon -Constraints Decomposition by Yaohong Yang, Sammie Katt, Samuel Kaski (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-15959.txt.

Description

Yang, Katt, and Kaski propose STAGE-BO, a multi-objective Bayesian optimization method for several costly goals at once. Existing methods struggle with coverage, scale, constraints, and preferences. By finding the Pareto point with the largest uncovered gap, the method sets adaptive limits that turn the problem into a sequence of smaller limited problems, solved by constrained expected improvement. It gives uniform Pareto spread without hypervolume math, and handles constraints and user wishes. Tests on synthetic and real-world benchmarks show better coverage and competitive hypervolume results.

Catalog record

CREATORYaohong Yang, Sammie Katt, Samuel Kaski
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE49562
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.15959

Browse the full library · About the Tower of Babel