Machine-Learning Surrogate Models for Predicting Electronic Structures


Machine-Learning Surrogate Models for Predicting Electronic Structures

Cangi, A.

The successful characterization of high energy density (HED) phenomena in laboratories using photon sources or pulsed power facilities is possible only with numerical modeling for design, diagnostic development, and data interpretation. The persistence of electron correlation is one of the greatest challenges for accurate numerical modeling and has hitherto impeded our ability to model HED phenomena across multiple length and time scales at sufficient accuracy. Standard methods from electronic structure theory capture electron correlation at high accuracy, but are limited to small scales due to their high computational cost.
Artificial intelligence (AI) has emerged as a powerful tool for analyzing complex datasets. It has the potential to accelerate electronic structure calculations to hitherto unattainable scales [1].
In this talk, I will present our recent efforts on devising a data-driven and physics-informed machine-learning workflow to tackle this challenge. Based on first-principles data we generate machine-learning surrogate models that replace traditional density functional theory calculations. Our Materials Learning Algorithms framework [2] predicts the electronic structure and related properties of matter under extreme conditions highly efficiently while maintaining the accuracy of traditional methods [3]. Our most recent results towards automated machine-learning save orders of magnitude in computational efforts for finding suitable neural network models and set the stage for large-scale investigations based on AI-driven methods [4].

References:

[1] L. Fiedler, K. Shah, M. Bussmann, A. Cangi, A Deep Dive into Machine Learning Density Functional Theory for Materials Science and Chemistry, Phys. Rev. Materials 6, 040301, (2022).
[2] A. Cangi, J. A. Ellis, L. Fiedler, D. Kotik, N. A. Modine, V. Oles, G. A. Popoola, S. Rajamanickam, S. Schmerler, J. A. Stephens, A. P. Thompson, MALA (Version 0.2.0), https://doi.org/10.5281/zenodo.5557254 (2021).
[3] J. A. Ellis, L. Fiedler, G. A. Popoola, N. A. Modine, J. A. Stephens, A. P. Thompson, A. Cangi, S. Rajamanickam, Phys. Rev. B 104, 035120 (2021).
[4] L. Fiedler, N. Hoffmann, P. Mohammed, G. A. Popoola, T. Yovell, V. Oles, J. A. Ellis, S. Rajamanickam, A. Cangi, arXiv:2202.09186 (2022).

Keywords: Quantum mechanics; Electronic structure theory; Density functional theory; Machine learning; Neural networks

  • Poster
    Advancing Quantum Mechanics with Mathematics and Statistics, Workshop IV: Monte Carlo and Machine Learning Approaches in Quantum Mechanics, 23.-27.05.2022, University of California, Los Angeles, United States
  • Lecture (others)
    Invitation to the Department of Chemistry, University of California, Irvine, 19.05.2022, Irvine, CA, United States

Permalink: https://www.hzdr.de/publications/Publ-35447
Publ.-Id: 35447