Exploring model complexity in machine learned potentials for simulated properties
Authors:
Open Access
- 18 September 2023
- journal article
- research article
- Published by Springer Nature in Journal of Materials Research
- Vol. 38 (24) , 5136-5150
Abstract
Machine learning (ML) enables the development of interatomic potentials with the accuracy of first principles methods while retaining the speed and parallel efficiency of empirical potentials. While ML potentials traditionally use atom-centered descriptors as inputs, different models such as linear regression and neural networks map descriptors to atomic energies and forces. This begs the question: what is the improvement in accuracy due to model complexity irrespective of descriptors? We curate three datasets to investigate this question in terms of ab initio energy and force errors: (1) solid and liquid silicon, (2) gallium nitride, and (3) the superionic conductor Li $$_{10}$$ Ge(PS $$_{6}$$ ) $$_{2}$$ (LGPS). We further investigate how these errors affect simulated properties and verify if the improvement in fitting errors corresponds to measurable improvement in property prediction. By assessing different models, we observe correlations between fitting quantity (e.g. atomic force) error and simulated property error with respect to ab initio values. Graphical abstract
Keywords
Funding Information
- Fusion Energy Sciences (20-023149)
- U.S. Air Force (FA9550-22-1-035)
- Office of Science (17-SC- 20-SC)
- U.S. Department of Energy (DE-NA0003965)
- Multidisciplinary University Research Initiative (00014-18-1-2429)
This publication has 42 references indexed in Scilit:
- Parallel Multistream Training of High-Dimensional Neural Network PotentialsJournal of Chemical Theory and Computation, 2019
- Atomic cluster expansion for accurate and transferable interatomic potentialsPhysical Review B, 2019
- Energy-based descriptors to rapidly predict hydrogen storage in metal–organic frameworksMolecular Systems Design & Engineering, 2018
- Extending the accuracy of the SNAP interatomic potential formThe Journal of Chemical Physics, 2018
- Ultrawide‐Bandgap Semiconductors: Research Opportunities and ChallengesAdvanced Electronic Materials, 2017
- Spectral neighbor analysis method for automated generation of quantum-accurate interatomic potentialsJournal of Computational Physics, 2015
- On representing chemical environmentsPhysical Review B, 2013
- Generalized Neural-Network Representation of High-Dimensional Potential-Energy SurfacesPhysical Review Letters, 2007
- Multilayer feedforward networks are universal approximatorsNeural Networks, 1989
- Superionic conductors: Transitions, structures, dynamicsPhysics Reports, 1979