The Bigger the Better? Accurate Molecular Potential Energy Surfaces from Minimalist Neural Networks

TL;DR

This paper introduces KerNN, a minimalist neural network approach for molecular potential energy surfaces that reduces parameters, accelerates computation, and enhances extrapolation, maintaining high accuracy for spectroscopy and reaction dynamics.

Contribution

KerNN combines kernel methods with neural networks to create a more efficient and extrapolation-capable PES model with fewer parameters than existing approaches.

Findings

KerNN significantly reduces training and evaluation times.

KerNN maintains high accuracy comparable to state-of-the-art models.

KerNN demonstrates superior extrapolation capabilities beyond training data.

Abstract

Atomistic simulations are a powerful tool for studying the dynamics of molecules, proteins, and materials on wide time and length scales. Their reliability and predictiveness, however, depend directly on the accuracy of the underlying potential energy surface (PES). Guided by the principle of parsimony this work introduces KerNN, a combined kernel/neural network-based approach to represent molecular PESs. Compared to state-of-the-art neural network PESs the number of learnable parameters of KerNN is significantly reduced. This speeds up training and evaluation times by several orders of magnitude while retaining high prediction accuracy. Importantly, using kernels as the features also improves the extrapolation capabilities of KerNN far beyond the coverage provided by the training data which solves a general problem of NN-based PESs. KerNN applied to spectroscopy and reaction dynamics shows excellent performance on test set statistics and observables including vibrational bands computed from classical and quantum simulations.