Machine Learning Directed Optimization of Classical Molecular Modeling Force Fields
Machine Learning Directed Optimization of Classical Molecular Modeling Force Fields
复制标题
DOI：
10.1021/acs.jcim.1c00448
复制
发表时间：
2021-08-17
影响因子：
5.6
通讯作者：
Maginn, Edward J.
中科院分区：
文献类型：
--
作者：
Befort, Bridgette J.;DeFever, Ryan S.;Maginn, Edward J.
精确的力场是预测性分子模拟所必需的。然而，开发精确再现实验性质的力场是具有挑战性的。在这里，我们提出了一个用于力场参数化的机器学习指导的多目标优化工作流程，该工作流程评估了数百万个预期的力场参数集，同时只需要其中一小部分进行分子模拟测试。我们证明了该方法的一般性，并确定多个低误差参数集两个不同的测试用例：模拟氢氟碳化合物（HFC）的汽液平衡（VLE）和高氯酸铵（AP）的结晶相。我们讨论了我们的力场优化工作流程的挑战和影响。
Accurate force fields are necessary for predictive molecular simulations. However, developing force fields that accurately reproduce experimental properties is challenging. Here, we present a machine learning directed, multiobjective optimization workflow for force field parametrization that evaluates millions of prospective force field parameter sets while requiring only a small fraction of them to be tested with molecular simulations. We demonstrate the generality of the approach and identify multiple low-error parameter sets for two distinct test cases: simulations of hydrofluorocarbon (HFC) vapor-liquid equilibrium (VLE) and an ammonium perchlorate (AP) crystal phase. We discuss the challenges and implications of our force field optimization workflow.