Bayesian optimization for high-dimensional coarse-grained model parameterization: a case study on Pebax polymer
Author(s)
Type
Journal Article
Abstract
Coarse-grained (CG) force field models are extensively utilized in material simulations because of their scalability. Ordinarily, these models are parametrized using hybrid strategies that sequentially integrate top-down and bottom-up approaches. However, this combination restricts the capacity to jointly optimize all parameters. Although Bayesian optimization (BO) has been explored as an alternative search strategy to identify well-optimized CG parameters, its application has conventionally been limited to low-dimensional scenarios. This has contributed to the assumption that BO is unsuitable for more complex CG models, which often involve a large number of parameters. In this study, we challenge this assumption by successfully extending BO, using the tree-structured Parzen estimator (TPE) model, to optimize a high-dimensional CG model. Specifically, we show that a 41-parameter CG model of Pebax-1657, a copolymer composed of alternating polyamide and polyether segments, can be effectively parametrized using BO, resulting in a model that accurately reproduces the key physical properties of its parent atomistic representation. Our optimization framework simultaneously targets structural and thermodynamic properties, namely, density, radius of gyration, and glass transition temperature. Compared to traditional search algorithms, BO-TPE not only converges faster but also delivers consistent improvements over more standard parametrization approaches.
Date Issued
2026-03-10
Date Acceptance
2026-01-02
Citation
Journal of Chemical Theory and Computation, 2026, 22 (5), pp.2358-2368
ISSN
1549-9618
Publisher
American Chemical Society
Start Page
2358
End Page
2368
Journal / Book Title
Journal of Chemical Theory and Computation
Volume
22
Issue
5
Copyright Statement
Copyright © 2026 The Authors. Published by American Chemical Society. This publication is licensed under CC-BY 4.0 .
License URL
Identifier
10.1021/acs.jctc.5c01500
Publication Status
Published
Date Publish Online
2026-01-21