Machine Learning Force Field Parameters from Ab Initio Data

Abstract

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.

ElectroLens: Understanding Atomistic Simulations through Spatially-Resolved Visualization of High-Dimensional Features

“…Future work includes implementing other projections into ElectroLens, as well as studying methods to automatically infer informative feature combinations. We also plan to focus on applications of the tool to the increasingly popular NN force-fields for molecular dynamics simulations [6,10,18,21], including improved handling of time-dependent data sets. Further, we expect that ElectroLens may be useful for a wider set of problems in chemistry, physics, and engineering involving spatially-resolved high-dimensional data.…”

First-Principles Many-Body Nonadditive Polarization Energies from Monomer and Dimer Calculations Only: A Case Study on Water

“…The (H 2 O) 16 isomers have reference energies within only 6 kJ mol −1 , which is half the energy range of the water hexamers. The 4444-a and 4444-b isomers differ in energy by only 1.2 kJ mol −1 and the boat-b and anti-boat isomers differ by 1.4 kJ mol −1 .…”

“…The (H 2 O) 16 clusters have been optimized by Yoo and Xantheas [69] and this set includes two bonding variants of the "4444" structure and two of the "boat" structure. The fifth structure, the "anti-boat", was estimated [69] to have an energy lying between the two boat structures.…”

“…In this way, appropriate levels of theory and basis sets can be used for each level, and numerical issues like the basis-set superposition error (BSSE) can be corrected. The intermolecular interaction energies for the (H 2 O) 16 isomers are given in Table III and are terms including 2-body to 4-body contributions, E int [2B-4BB-,B] are displayed in Figure 17. Note that SAMBA reference energies are not available for the boat-a isomer.…”

“…In Figure 18 we display the many-body energy decomposition of the (H 2 O) 16 clusters. As with the hexamers (see Figure 14) we see that the excellent total interaction energies from the DIFF models results from a cancellation of errors made in the 2-body and many-body energies.…”