Atomic positions independent descriptor for machine learning of material properties

TL;DR

This paper introduces an atomic-position independent descriptor for machine learning of material properties, enabling accurate predictions without requiring atomic positions, thus facilitating rapid screening of vast material spaces.

Contribution

The authors propose a novel descriptor based solely on crystallographic symmetry and atomic numbers, removing the need for atomic positions in ML models for materials.

Findings

Achieved 0.07 eV/atom MAE in formation energy predictions.

Enabled high-accuracy predictions using only symmetry and atomic numbers.

Facilitated large-scale materials screening with simplified descriptors.

Abstract

The high-throughput screening of periodic inorganic solids using machine learning methods requires atomic positions to encode structural and compositional details into appropriate material descriptors. These atomic positions are not available {\it a priori} for new materials which severely limits exploration of novel materials. We overcome this limitation by using only crystallographic symmetry information in the structural description of materials. We show that for materials with identical structural symmetry, machine learning is trivial and accuracies similar to that of density functional theory calculations can be achieved by using only atomic numbers in the material description. For machine learning of formation energies of bulk crystalline solids, this simple material descriptor is able to achieve prediction mean absolute errors of only 0.07 eV/atom on a test dataset consisting of more than 85,000 diverse materials. This atomic-position independent material descriptor presents a new route of materials discovery wherein millions of materials can be screened by training a machine learning model over a drastically reduced subspace of materials.