Gaussian process regression for geometry optimization
Gaussian process regression for geometry optimization
复制标题
DOI：
10.1063/1.5017103
复制
发表时间：
2018-03-07
影响因子：
4.4
通讯作者：
Kaestner, Johannes
中科院分区：
文献类型：
--
作者：
Denzel, Alexander;Kaestner, Johannes
我们实现了一个基于高斯过程回归（GPR）的几何优化器来寻找势能表面上的最小结构。我们测试了母核的二次可微形式和指数核的平方。Matern内核的性能要好得多。我们给出了优化过程的详细描述。这些包括为了获得更高程度的内插与外推而由GPR产生的超调步长。在针对DL-FIND库的有限内存Broyden-Fletcher-Goldfarb-Shanno优化器在26个测试系统上的基准测试中，我们发现新的优化器通常可以减少所需优化步骤的数量。AIP出版社出版。
We implemented a geometry optimizer based on Gaussian process regression (GPR) to find minimum structures on potential energy surfaces. We tested both a two times differentiable form of the Matern kernel and the squared exponential kernel. The Matern kernel performs much better. We give a detailed description of the optimization procedures. These include overshooting the step resulting from GPR in order to obtain a higher degree of interpolation vs. extrapolation. In a benchmark against the Limited-memory Broyden-Fletcher-Goldfarb-Shanno optimizer of the DL-FIND library on 26 test systems, we found the new optimizer to generally reduce the number of required optimization steps. Published by AIP Publishing.