Machine learning for predictive condensed-phase simulation

TL;DR

This paper demonstrates how Bayesian machine learning can enhance the accuracy of molecular simulations, especially for water, by correcting DFT calculations to better predict energies, structures, and dynamics.

Contribution

The authors introduce a Bayesian machine learning approach to systematically improve DFT accuracy for molecular materials like water.

Findings

Improved energy predictions for water clusters and ice structures.

Enhanced modeling of liquid water's structure and dynamics.

Systematic correction method for DFT using machine learning.

Abstract

We show how machine learning techniques based on Bayesian inference can be used to reach new levels of realism in the computer simulation of molecular materials, focusing here on water. We train our machine-learning algorithm using accurate, correlated quantum chemistry, and predict energies and forces in molecular aggregates ranging from clusters to solid and liquid phases. The widely used electronic-structure methods based on density-functional theory (DFT) give poor accuracy for molecular materials like water, and we show how our techniques can be used to generate systematically improvable corrections to DFT. The resulting corrected DFT scheme gives remarkably accurate predictions for the relative energies of small water clusters and of different ice structures, and greatly improves the description of the structure and dynamics of liquid water.