📊 View usage/download statistics
Bgolearn is a flexible and extensible Python package for Bayesian Global Optimization (BGO). It is specifically designed to accelerate materials discovery via active learning and adaptive sampling strategies.
- 🧠 Bayesian Optimization Core: Supports single- and multi-objective optimization using GPR-based surrogate models.
- 🧪 Materials Design-Oriented: Tailored for high-throughput experiments and structure–property optimization workflows.
- 🔁 Active Learning Framework: Combines uncertainty sampling and exploration–exploitation balance strategies.
- 🎯 Customizable Acquisition Functions: Includes EI, PI, UCB, and supports user-defined strategies.
- 🌐 User Interface + Web Deployment: Works with BgoFace for intuitive web-based control.
- 📺 Code Walkthrough: Watch on BiliBili
- 🧪 Sample Code + Datasets: CodeDemo Repository
| Name | Description |
|---|---|
| 🔗 Bgolearn | Core source code of the Bayesian Global Optimization framework |
| 🔗 MultiBgolearn | Extension for multi-objective optimization |
| 🔗 BgoFace | Graphical user interface (GUI) for interactive BGO |
| 🔗 CodeDemo | Example scripts and synthetic datasets |
| 🔗 Document | Official documentation site |
| 🔗 MLMD | A programming-free platform for ML-based materials design |
If you use Bgolearn in your research, please cite:
Cao, B., Su, T., Yu, S., Li, T., Zhang, T., Zhang, J., ... & Zhang, T. Y. (2024). Active learning accelerates the discovery of high strength and high ductility lead-free solder alloys. Materials & Design, 241, 112921. https://doi.org/10.1016/j.matdes.2024.112921
Explore more works using Bgolearn on Google Scholar
We welcome contributions and suggestions! Please ⭐️ the repo Bgolearn if you find it helpful.
