We have encountered instances where papers have used the Bgolearn package but failed to cite it. We will continue to monitor such cases and will contact the respective journals to request proper citation.

View package usage statistics / download counts

• Bgolearn : The Source Code of Bgolearn
• MultiBgolearn: Multi-Objective Bayesian Global Optimization for Materials Design
• BgoFace : User interface of the Bgolearn platform
• CodeDemo : Provides code demonstrations and data to illustrate the application of Bgolearn
• Document : The Document of Bgolearn
• MLMD : A programming-free AI platform to predict and design materials

• 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

Multi-target selection is not the same as multi-target optimization!

• Multi-target optimization considers multiple properties simultaneously in both prediction and utility space. It can be implemented using MultiBgolearn in Python.

• Multi-target selection typically considers two properties independently and then combines them in either the property space or utility space using a Pareto front. A method is constructed to select one solution from the points on the Pareto front. This can be implemented using BgoKit in Python.

Note: We are not claiming that one approach is better than the other; they are fundamentally different. Multi-target optimization truly accounts for the interdependencies between properties, whereas multi-target selection treats properties more independently.

If you are still unclear about the differences, please refer to the video for further explanation.