Publications Repository - Helmholtz-Zentrum Dresden-Rossendorf

1 Publication

Investigation of transferability in LDOS based DFT surrogate models for multiscale simulations

Fiedler, L.; Cangi, A.

Density Functional Theory (DFT) is one of the most important computational tools for materials science, as it combines high accuracy with general computational feasibility. However, applications important to scientific progress can pose problems to even the most advanced and efficient DFT codes due to size and/or complexity of the underlying simulations. Namely the modeling of materials across multiple length and time scales at ambient or extreme conditions, necessary for the understanding of important physical phenomena such as radiation damages in fusion reactor walls, evade traditional ab-initio treatment.
DFT surrogate models are a useful tool in achieving this goal by reproducing DFT results at drastically reduced computational cost by using machine learning methods. Yet, a lack of transferability of many approaches lead to repeated and costly training data generation procedures. Here, we present results of an investigation to transfer such machine learning DFT surrogate models between different simulation cell sizes, with the goal of reducing the overall amount of computational time for training data generation. The models are based upon the Materials Learning Algorithms (MALA) package [1] and the therein implemented LDOS based machine learning workflow [2].
[2]: J. A. Ellis et al., Phys. Rev. B 104, 035120, 2021

Keywords: Machine Learning; Density Functional Theory; Surrogate Model

  • Lecture (Conference) (Online presentation)
    84. Jahrestagung der DPG und DPG-Tagung der Sektion Kondensierte Materie (SKM), 27.09.-01.10.2021, online, Deutschland

Publ.-Id: 33295