Witt William C, van der Oord Cas, Gelžinytė Elena, Järvinen Teemu, Ross Andres, Darby James P, Ho Cheuk Hin, Baldwin William J, Sachs Matthias, Kermode James, Bernstein Noam, Csányi Gábor, Ortner Christoph
Department of Materials Science and Metallurgy, University of Cambridge, Cambridge, United Kingdom.
Engineering Laboratory, University of Cambridge, Cambridge CB2 1PZ, United Kingdom.
J Chem Phys. 2023 Oct 28;159(16). doi: 10.1063/5.0158783.
We introduce ACEpotentials.jl, a Julia-language software package that constructs interatomic potentials from quantum mechanical reference data using the Atomic Cluster Expansion [R. Drautz, Phys. Rev. B 99, 014104 (2019)]. As the latter provides a complete description of atomic environments, including invariance to overall translation and rotation as well as permutation of like atoms, the resulting potentials are systematically improvable and data efficient. Furthermore, the descriptor's expressiveness enables use of a linear model, facilitating rapid evaluation and straightforward application of Bayesian techniques for active learning. We summarize the capabilities of ACEpotentials.jl and demonstrate its strengths (simplicity, interpretability, robustness, performance) on a selection of prototypical atomistic modelling workflows.
我们介绍了ACEpotentials.jl,这是一个用Julia语言编写的软件包,它使用原子团簇展开方法[R. Drautz,《物理评论B》99,014104(2019年)]从量子力学参考数据构建原子间势。由于原子团簇展开方法提供了对原子环境的完整描述,包括对整体平移和旋转以及同类原子置换的不变性,因此得到的势可以系统地改进且数据效率高。此外,描述符的表现力使得能够使用线性模型,便于快速评估和直接应用贝叶斯技术进行主动学习。我们总结了ACEpotentials.jl的功能,并在一些典型的原子建模工作流程中展示了它的优势(简单性、可解释性、稳健性、性能)。