A neural network potential with self-trained atomic fingerprints: a test with the mW water potential

TL;DR

This paper introduces a neural network potential utilizing self-trained atomic fingerprints, effectively modeling water's complex behaviors across various phases and conditions, with improved accuracy through an annealing training protocol.

Contribution

It develops a novel neural network potential with self-trained atomic fingerprints and an annealing training method, enhancing modeling accuracy for water across diverse states.

Findings

Accurately reproduces mW water model over wide density and temperature ranges.

Captures phase transitions and structural properties not explicitly trained on.

Improves training stability and accuracy with annealing protocol.

Abstract

We present a neural network (NN) potential based on a new set of atomic fingerprints built upon two- and three-body contributions that probe distances and local orientational order respectively. Compared to existing NN potentials, the atomic fingerprints depend on a small set of tuneable parameters which are trained together with the neural network weights. To tackle the simultaneous training of the atomic fingerprint parameters and neural network weights we adopt an annealing protocol that progressively cycles the learning rate, significantly improving the accuracy of the NN potential. We test the performance of the network potential against the mW model of water, which is a classical three-body potential that well captures the anomalies of the liquid phase. Trained on just three state points, the NN potential is able to reproduce the mW model in a very wide range of densities and temperatures, from negative pressures to several GPa, capturing the transition from an open random tetrahedral network to a dense interpenetrated network. The NN potential also reproduces very well properties for which it was not explicitly trained, such as dynamical properties and the structure of the stable crystalline phases of mW.