A Differentiable Neural-Network Force Field for Ionic Liquids

TL;DR

NeuralIL is a differentiable neural network model that accurately predicts the potential energy and forces of ionic liquids, enabling efficient simulations with minimal training data and no need for explicit long-range interaction modeling.

Contribution

The paper introduces NeuralIL, a fully differentiable neural network force field for ionic liquids that leverages spherical Bessel descriptors and force training to improve accuracy and efficiency.

Findings

Achieves <0.05 kcal/mol energy accuracy on ethylammonium nitrate

Requires fewer atomic configurations due to force-based training

Does not need explicit long-range interaction modeling

Abstract

We present NeuralIL, a model for the potential energy of an ionic liquid that accurately reproduces first-principles results with orders-of-magnitude savings in computational cost. Based on a multilayer perceptron and spherical Bessel descriptors of the atomic environments, NeuralIL is implemented in such a way as to be fully automatically differentiable. It can thus be trained on ab-initio forces instead of just energies, to make the most out of the available data, and can efficiently predict arbitrary derivatives of the potential energy. Using ethylammonium nitrate as the test system, we obtain out-of-sample accuracies better than 2 meV/atom (<0.05 kcal/mol) in the energies and 70 meV/{\AA} in the forces. We show that encoding the element specific density in the spherical Bessel descriptors is key to achieving this. Harnessing the information provided by the forces drastically reduces the amount of atomic configurations required to train a neural network force field based on atom-centered descriptors. We choose the Swish-1 activation function and discuss the role of this choice in keeping the neural network differentiable. Furthermore, the possibility of training on small data sets allows for an ensemble-learning approach to the detection of extrapolation. Finally, we find that a separate treatment of long-range interactions is not required to achieve a high-quality representation of the potential energy surface of these dense ionic systems.