A universal model for the formation energy prediction of inorganic compounds

TL;DR

This paper presents a universal machine learning model trained on a large dataset of inorganic compounds to accurately predict formation energies, aiding rapid materials discovery.

Contribution

The study introduces a highly accurate, universally applicable machine learning model for formation energy prediction using novel structure-dependent descriptors.

Findings

Achieved R2=0.982 and MAE=0.07 eV/atom in predictions

Model effectively covers a large phase space of inorganic materials

Proposed descriptors incorporate electronegativity and structural information

Abstract

Harnessing the recent advance in data science and materials science, it is feasible today to build predictive models for materials properties. In this study, we employ the data of high-throughput quantum mechanics calculations based on 170,714 inorganic crystalline compounds to train a machine learning model for formation energy prediction. Different from the previous work, our model reaches a fairly good predictive ability (R2=0.982 and MAE=0.07 eVatom-1, DenseNet model) and meanwhile can be universally applied to the large phase space of inorganic materials. The improvement comes from several effective structure-dependent descriptors that are proposed to take the information of electronegativity and structure into account. This model can provide a useful tool to search for new materials in a vast phase space in a fast and cost-effective manner.