Bgolearn: a unified Bayesian optimization framework for accelerating materials discovery
11 affiliations
- 1Department of Physics, City University of Hong Kong, Hong Kong SAR, China
- 2Guangzhou Municipal Key Laboratory of Materials Informatics, Advanced Materials Thrust, The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China
- 3Materials Genome Institute, Shanghai University, Shanghai, China
- 4State Key Laboratory of Materials for Advanced Nuclear Energy, Shanghai University, Shanghai, China
- 5School of Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai, China
- 6School of Aerospace, Mechanical and Mechatronic Engineering, School of Computer Science, The University of Sydney, Sydney, NSW, Australia
- 7School of Materials Science and Engineering, Harbin Institute of Technology (Shenzhen), Shenzhen, China
- 8University College London, London, UK
- 9Suzhou Laboratory, Suzhou, China
- 10State Key Laboratory for Mechanical Behavior of Materials, Xi’an Jiaotong University, Xi’an, China
- 11AiMaterials Research, Santa Fe, NM, USA
Abstract
Efficient exploration of vast compositional and processing spaces is essential for accelerated materials discovery. Bayesian optimization (BO) provides a principled strategy for identifying optimal materials with minimal experiments, yet its adoption in materials science ishindered by implementation complexity and limited domain-specific tools. Here, we presentBgolearn, a comprehensive Python framework that makes BO accessible and practical for materials research through an intuitive interface, robust algorithms, and materials-oriented workflows. Bgolearn supports both single-objective and multi-objective Bayesian optimization withmultiple acquisition functions (e.g., expected improvement, upper confidence bound, probability of improvement, and expected hypervolume improvement etc.), diverse surrogate models(including Gaussian processes, random forests, and gradient boosting etc.), and bootstrapbased uncertainty quantification. Benchmark studies show that Bgolearn reduces the numberof required experiments by 40–60% compared with random search, grid search, and geneticalgorithms, while maintaining comparable or superior solution quality. Its effectiveness isdemonstrated not only through the studies presented in this paper, such as the identificationof maximum-elastic-modulus triply periodic minimal surface structures, ultra-high-hardnesshigh-entropy alloys, and high-strength, high-ductility medium-Mn steels, but also by numerous publications that have proven its impact in material discovery. With a modular architecture that integrates seamlessly into existing materials workflows and a graphical user interface (BgoFace) that removes programming barriers, Bgolearn establishes a practical andreliable platform for Bayesian optimization in materials science, and is openly available athttps://github.com/Bin-Cao/Bgolearn.
Related
- Accelerating the discovery of nanoporous gas separation materials via Coulomb-enhanced classical density functional theory
- Accelerating Advanced Materials Innovation with New Approach Methodologies
- Accelerating Quantum Simulations of Materials Through Parameter and Ansatz Transfer Strategies
- Accelerating Catalyst Discovery with Artificial Intelligence Using Molecular Dual-Atom Site Catalysts as a Case Study
- Accelerating discovery of phosphorescent blue OLED emitters using chemical foundation models
BibTeX
@article{cao2026bgolearn,
title = {{Bgolearn: a unified Bayesian optimization framework for accelerating materials discovery}},
author = {Bin Cao and Jie Xiong and Jiaxuan Ma and Yuan Tian and Yirui Hu and Mengwei He and Longhan Zhang and Jiayu Wang and Jian Hui and Li Liu and Dezhen Xue and Turab Lookman and Jun Wang and Tong‐Yi Zhang},
journal = {npj Computational Materials},
year = {2026},
month = jul,
doi = {10.1038/s41524-026-02226-3},
url = {https://doi.org/10.1038/s41524-026-02226-3},
note = {CC BY-NC-ND},
}
RIS
TY - JOUR
TI - Bgolearn: a unified Bayesian optimization framework for accelerating materials discovery
AU - Bin Cao
AU - Jie Xiong
AU - Jiaxuan Ma
AU - Yuan Tian
AU - Yirui Hu
AU - Mengwei He
AU - Longhan Zhang
AU - Jiayu Wang
AU - Jian Hui
AU - Li Liu
AU - Dezhen Xue
AU - Turab Lookman
AU - Jun Wang
AU - Tong‐Yi Zhang
PY - 2026
DA - 2026/07/14
JO - npj Computational Materials
DO - 10.1038/s41524-026-02226-3
UR - https://doi.org/10.1038/s41524-026-02226-3
AB - Efficient exploration of vast compositional and processing spaces is essential for accelerated materials discovery. Bayesian optimization (BO) provides a principled strategy for identifying optimal materials with minimal experiments, yet its adoption in materials science ishindered by implementation complexity and limited domain-specific tools. Here, we presentBgolearn, a comprehensive Python framework that makes BO accessible and practical for materials research through an intuitive interface, robust algorithms, and materials-oriented workflows. Bgolearn supports both single-objective and multi-objective Bayesian optimization withmultiple acquisition functions (e.g., expected improvement, upper confidence bound, probability of improvement, and expected hypervolume improvement etc.), diverse surrogate models(including Gaussian processes, random forests, and gradient boosting etc.), and bootstrapbased uncertainty quantification. Benchmark studies show that Bgolearn reduces the numberof required experiments by 40–60% compared with random search, grid search, and geneticalgorithms, while maintaining comparable or superior solution quality. Its effectiveness isdemonstrated not only through the studies presented in this paper, such as the identificationof maximum-elastic-modulus triply periodic minimal surface structures, ultra-high-hardnesshigh-entropy alloys, and high-strength, high-ductility medium-Mn steels, but also by numerous publications that have proven its impact in material discovery. With a modular architecture that integrates seamlessly into existing materials workflows and a graphical user interface (BgoFace) that removes programming barriers, Bgolearn establishes a practical andreliable platform for Bayesian optimization in materials science, and is openly available athttps://github.com/Bin-Cao/Bgolearn.
SN - 2057-3960
ER -