A Transferable Active-Learning Strategy for Reactive Molecular Force Fields

Abstract

<p>Predictive simulations of dynamic processes in molecular systems require fast, accurate and reactive interatomic potentials. Machine learning offers a promising approach to construct force-field models for large-scale molecular simulation by fitting to high-level quantum-mechanical data. However, machine-learned force fields generally require considerable human intervention and data volume. Here we show that, by leveraging hierarchical and active learning, accurate Gaussian Approximation Potential (GAP) models for diverse chemical systems can be developed in an autonomous way, requiring only hundreds to a few thousand energy and gradient evaluations on the reference potential-energy surface. Our approach relies on a decomposition of the condensed-phase molecular system into intra- and inter-molecular terms, and on the definition of a prospective error metric to quantify accuracy. We demonstrate applications to a range of molecular systems: from bulk water, organic solvents, and a solvated ion onwards to the description of chemical reactivity, including, a bifurcating Diels–Alder reaction in the gas phase and non-equilibrium dynamics (S<sub>N</sub>2 reaction) in explicit solvent. The method provides a route to routinely generating machine-learned force fields for complex and/or reactive molecular systems. </p>

Gaussian Process Regression for Materials and Molecules

“…One of the promises of ML force fields for molecules is that they will enable the accurate and routine construction of general reactive molecular force fields. 420 There is scant research on this as yet, and Figure 29 b shows that in comparison with closed shell molecules (such as those in QM9), describing open-shell radicals is much harder: the errors of the SOAP-based kernel model are three times larger on the Rad-6 dataset, which consists of all closed- and open-shell molecules containing C, H, and O with up to six non-hydrogen atoms.…”

Linear Atomic Cluster Expansion Force Fields for Organic Molecules: beyond RMSE

“…The insufficient nature of mean error metrics has been pointed out before. [37][38][39] In addition to the above data sets, we also demonstrate the use of ACE on a slightly larger, significantly more flexible molecule that is more representative of the needs of medicinal chemistry applications.…”

Linear Atomic Cluster Expansion Force Fields for Organic Molecules: beyond RMSE

“…The insufficient nature of mean error metrics has been pointed out before. [37][38][39] In addition to the above data sets, we also demonstrate the use of ACE on a slightly larger, significantly more flexible molecule that is more representative of the needs of medicinal chemistry applications.…”