Building machine learning force fields for nanoclusters

Abstract

We assess Gaussian process (GP) regression as a technique to model interatomic forces in metal nanoclusters by analysing the performance of 2-body, 3-body and many-body kernel functions on a set of 19-atom Ni cluster structures. We find that 2-body GP kernels fail to provide faithful force estimates, despite succeeding in bulk Ni systems. However, both 3- and many-body kernels predict forces within a $\sim$0.1 eV/$\text{A˚}$ average error even for small training datasets, and achieve high accuracy even on out-of-sample, high temperature, structures. While training and testing on the same structure always provides satisfactory accuracy, cross-testing on dissimilar structures leads to higher prediction errors, posing an extrapolation problem. This can be cured using heterogeneous training on databases that contain more than one structure, which results in a good trade-off between versatility and overall accuracy. Starting from a 3-body kernel trained this way, we build an efficient non-parametric 3-body force field that allows accurate prediction of structural properties at finite temperatures, following a newly developed scheme [Glielmo et al. PRB 97, 184307 (2018)]. We use this to assess the thermal stability of Ni${}_{19}$ nanoclusters at a fractional cost of full ab initio calculations.

arXiv comment14 pages, 14 figures

References in corpus (9)

• Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
• Moment Tensor Potentials: a class of systematically improvable interatomic potentials
• Machine-learning based interatomic potential for amorphous carbon
• Accurate Interatomic Force Fields via Machine Learning with Covariant Kernels
• How to quantify energy landscapes of solids
• Efficient nonparametric $n$-body force fields from machine learning
• Efficient non-parametric fitting of potential energy surfaces for polyatomic molecules with Gaussian processes
• Machine Learning for Atomic Forces in a Crystalline Solid: Transferability to Various Temperatures
• Al20+ does melt, albeit above the bulk melting temperature of aluminium

Cited by in corpus (12)

• A Deep Dive into Machine Learning Density Functional Theory for Materials Science and Chemistry
• Guest Editorial: Special Topic on Data-enabled Theoretical Chemistry
• Data-driven simulation and characterisation of gold nanoparticle melting
• PANNA: Properties from Artificial Neural Network Architectures
• Fast Neural Network Approach for Direct Covariant Forces Prediction in Complex Multi-Element Extended Systems
• On Machine Learning Force Fields for Metallic Nanoparticles
• Compact atomic descriptors enable accurate predictions via linear models
• Neural Networks Potential from the Bispectrum Component: A Case Study on Crystalline Silicon
• Infrared spectra of neutral polycyclic aromatic hydrocarbons by machine learning
• Super-resolution in Molecular Dynamics Trajectory Reconstruction with Bi-Directional Neural Networks
• Building nonparametric $n$-body force fields using Gaussian process regression
• Critical assessment of machine-learned repulsive potentials for the Density Functional based Tight-Binding method: a case study for pure silicon