Jump to main content
Jump to site search


Addressing uncertainty in atomistic machine learning

Author affiliations

Abstract

Machine-learning regression has been demonstrated to precisely emulate the potential energy and forces that are output from more expensive electronic-structure calculations. However, to predict new regions of the potential energy surface, an assessment must be made of the credibility of the predictions. In this perspective, we address the types of errors that might arise in atomistic machine learning, the unique aspects of atomistic simulations that make machine-learning challenging, and highlight how uncertainty analysis can be used to assess the validity of machine-learning predictions. We suggest this will allow researchers to more fully use machine learning for the routine acceleration of large, high-accuracy, or extended-time simulations. In our demonstrations, we use a bootstrap ensemble of neural network-based calculators, and show that the width of the ensemble can provide an estimate of the uncertainty when the width is comparable to that in the training data. Intriguingly, we also show that the uncertainty can be localized to specific atoms in the simulation, which may offer hints for the generation of training data to strategically improve the machine-learned representation.

Graphical abstract: Addressing uncertainty in atomistic machine learning

Back to tab navigation
Please wait while Download options loads

Supplementary files

Publication details

The article was received on 17 Jan 2017, accepted on 29 Mar 2017 and first published on 18 Apr 2017


Article type: Perspective
DOI: 10.1039/C7CP00375G
Citation: Phys. Chem. Chem. Phys., 2017, Advance Article
  •   Request permissions

    Addressing uncertainty in atomistic machine learning

    A. A. Peterson, R. Christensen and A. Khorshidi, Phys. Chem. Chem. Phys., 2017, Advance Article , DOI: 10.1039/C7CP00375G

Search articles by author