Force Training Neural Network Potential Energy Surface Models

TL;DR

This paper develops a neural network potential energy surface model trained on both energies and forces, demonstrating improved accuracy with smaller datasets for a hydrogen transfer reaction, and highlights the importance of loss function weighting.

Contribution

It introduces a force training approach for neural network potentials that enhances accuracy and efficiency compared to energy-only training.

Findings

Force training significantly improves model accuracy with smaller datasets.

Proper weighting of force to energy in the loss function is crucial.

The model effectively captures the hydrogen transfer reaction dynamics.

Abstract

Machine learned chemical potentials have shown great promise as alternatives to conventional computational chemistry methods to represent the potential energy of a given atomic or molecular system as a function of its geometry. However, such potentials are only as good as the data they are trained on, and building a comprehensive training set can be a costly process. Therefore, it is important to extract as much information from training data as possible without further increasing the computational cost. One way to accomplish this is by training on molecular forces in addition to energies. This allows for three additional labels per atom within the molecule. Here we develop a neural network potential energy surface for studying a hydrogen transfer reaction between two conformers of C5H5. We show that, for a much smaller training set, force training can greatly improve the accuracy of the model compared to only training to energies. We also demonstrate the importance of choosing the proper force to energy weight ratio for the loss function to minimize the model test error.