Machine Learning Force Field Parameters from Ab Initio Data.
Machine Learning Force Field Parameters from Ab Initio Data.
复制标题
DOI：
10.1021/acs.jctc.7b00521
复制
发表时间：
2017-09-12
影响因子：
5.5
通讯作者：
Roux B
中科院分区：
文献类型：
--
作者：
Li Y;Li H;Pickard FC 4th;Narayanan B;Sen FG;Chan MKY;Sankaranarayanan SKRS;Brooks BR;Roux B
具有遗传算法（GA）的机器学习（ML）技术已经被应用于仅使用来自分子簇在MP2/6- 31 G（d，p）、DFP 2（fc）/jul-cc-pVDZ和DFP 2（fc）/jul-cc-pVTZ水平的量子力学（QM）计算的从头算数据来探索可极化力场参数，以预测实验凝聚相性质（即，密度和汽化热）。这种ML/GA方法的性能表现在4,943个二聚体的静电势和1,250个簇的相互作用能甲醇。QM计算的训练数据集与优化的力场模型之间可以实现良好的一致性。通过在机器学习过程中引入偏移因子来补偿QM计算的能量和由优化力场再现的能量的差异，实现了更好的结果，其中偏移因子保持QM能量表面的局部“形状”。在整个机器学习过程中，实验观测量不参与目标函数，而仅用于模型验证。在DFP 2（fc）/jul-cc-pVTZ水平下从QM数据优化的最佳模型似乎表现得甚至比原始AMOEBA力场（amoeba09.prm）更好，原始AMOEBA力场根据经验优化以匹配液体性质。目前的努力表明，使用机器学习技术开发描述性极化力场仅使用QM数据的可能性。ML/GA策略，以优化力场参数这里描述的可以很容易地扩展到其他分子系统。
Machine learning (ML) techniques with the genetic algorithm (GA) have been applied to explore a polarizable force field parameters using only ab initio data from quantum mechanics (QM) calculations of molecular clusters at the MP2/6-31G(d,p), DFMP2(fc)/jul-cc-pVDZ, and DFMP2(fc)/jul-cc-pVTZ levels to predict experimental condensed phase properties (i.e., density and heat of vaporization). The performance of this ML/GA approach is demonstrated on 4,943 dimers electrostatic potentials and 1,250 clusters interaction energies for methanol. Excellent agreement between the training dataset from QM calculations and the optimized force field model can be achieved. Better results are achieved by introducing an offset factor during the machine learning process to compensate for the discrepancy of the QM calculated energy and the energy reproduced by optimized force field, where the offset factor maintain the local “shape” of the QM energy surface. Throughout the machine learning process, experimental observables were not involved in the objective function, but were only used for model validation. The best model, optimized from the QM data at the DFMP2(fc)/jul-cc-pVTZ level, appears to perform even better than the original AMOEBA force field (amoeba09.prm), which was optimized empirically to match liquid properties. The present effort shows the possibility of using machine learning techniques to develop descriptive polarizable force field using only QM data. The ML/GA strategy to optimize force field parameters described here could easily be extended to other molecular systems.
登录
查看更多内容
影响因子：
4.4
作者：
Izvekov, S;Parrinello, M;Voth, GA
通讯作者：
Voth, GA
影响因子：
5.7
作者：
Cherukara, Mathew J.;Narayanan, Badri;Sankaranarayanan, Subramanian K. R. S.
通讯作者：
Sankaranarayanan, Subramanian K. R. S.
影响因子：
62.1
作者：
Lemkul JA;Huang J;Roux B;MacKerell AD Jr
通讯作者：
MacKerell AD Jr
影响因子：
4.4
作者：
CAMPBELL, ES;MEZEI, M
通讯作者：
MEZEI, M
影响因子：
1.2
作者：
Fliege, J;Svaiter, BF
通讯作者：
Svaiter, BF