Efficient and accurate machine-learning interpolation of atomic energies in compositions with many species

By Nongnuch Artrith, A. Urban, Gerbrand Ceder. 2017

235 citations recorded by the source corpus.

Author(s): Artrith, N; Urban, A; Ceder, G | Abstract: © 2017 us. Published by the American Physical Society. Machine-learning potentials (MLPs) for atomistic simulations are a promising alternative to conventional classical potentials. Current approaches rely on descriptors of the local atomic environment with dimensions that increase quadratically with the number of chemical species. In this paper, we demonstrate that such a scaling can be avoided in practice. We show that a mathematically simple and computationally efficient descriptor with constant complexity is sufficient to represent transition-metal oxide compositions and biomolecules containing 11 chemical species with a precision of around 3 meV/atom. This insight removes a perceived bound on the utility of MLPs and paves the way to investigate the physics of previously inaccessible materials with more than ten chemical species.

Subfields: Materials Chemistry, Physical and Theoretical Chemistry

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Materials Chemistry

In Physical and Theoretical Chemistry

Citations in the corpus

JAX, M.D.: End-to-End Differentiable, Hardware Accelerated, Molecular Dynamics in Pure Python (2019, background): “To address these issues, several projects developed adapters (Artrith & Urban, 2016; Artrith et al., 2017; Lot et al., 2019; Onat et al., 2018) between common ML languages, like Torch and Tensorflow, and common MD languages like LAMMPS.”
Emerging materials intelligence ecosystems propelled by machine learning (2020, background): “The generality of these potentials has been demonstrated for complex, multi-elemental transition-metal oxides and biomole-cules containing up to 11 chemical species 153 , which is far more than what can be easily targeted using traditional classical potentials.”
End-to-End Differentiable , Hardware Accelerated , Molecular Dynamics in Pure Python (2019, background): “To address these issues, several projects developed adapters (Artrith & Urban, 2016; Artrith et al., 2017; Lot et al., 2019; Onat et al., 2018) between common ML languages, like Torch and Tensorflow, and common MD languages like LAMMPS.”
Neuroevolution machine learning potentials: Combining high accuracy and low cost in atomistic simulations and application to heat transport (2021, unknown): “…the relative weights less important in systems with very diﬀerent atomic numbers, and then add the information of the central atom √ Z i to distinguish two conﬁgura-tions with the same environments but diﬀerent central atom species, which are indistinguishable in the previous approaches [42, 43].”