Machine-learned Multi-system Surrogate Models for Materials Prediction

Abstract

Surrogate machine-learning models are transforming computational materials science by predicting properties of materials with the accuracy of ab initio methods at a fraction of the computational cost. We demonstrate surrogate models that simultaneously interpolate energies of different materials on a dataset of 10 binary alloys (AgCu, AlFe, AlMg, AlNi, AlTi, CoNi, CuFe, CuNi, FeV, andNbNi) with 10 different species and all possible fcc, bcc, and hcp structures up to eight atoms in the unit cell, 15,950 structures intotal. We find that the deviation of prediction errors when increasing the number of simultaneously modeled alloys is <1 meV/atom. Several state-of-the-art materials representations and learning algorithms were found to qualitatively agree on the prediction errors of formation enthalpy with relative errors of <2.5% for all systems.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 18, 2019
Accession Number
AD1097843

Entities

People

  • Alexander V. Shapeev
  • Brayden Bekker
  • Chandramouli Nyshadham
  • Conrad W. Rosenbrock
  • David W. Wingate
  • Gus L. Hart
  • Gábor Csányi
  • Matthias Rupp
  • Tim Mueller

Organizations

  • Brigham Young University

Tags

Communities of Interest

  • Autonomy
  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Accuracy
  • Algorithms
  • Atoms
  • Bayesian Networks
  • Binary Alloys
  • Chemical Elements
  • Computational Science
  • Crystal Structure
  • Crystals
  • Deep Learning
  • Density Functional Theory
  • Information Science
  • Machine Learning
  • Materials
  • Materials Science
  • Neural Networks
  • Solid State Physics

Fields of Study

  • Physics

Readers

  • Computational Modeling and Simulation
  • Powder metallurgy of Titanium alloys.
  • Quantum Chemistry

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Neural Networks