Surrogate functionals for machine-learned orbital-free density functional theory
The Journal of Chemical Physics · 21.07.2026 · 165 (3)
DOI: 10.1063/5.0336653 ↗Kurz gefasst
Fachartikel in der Zeitschrift The Journal of Chemical Physics, online erschienen am 17.07.2026. Verfasst von Roman Remme und 1 weiteren Autor:innen. Beteiligte Einrichtung: Interdisciplinary Center for Scientific Computing (IWR), Heidelberg University, Heidelberg, Germany. Die Arbeit nennt 43 Literaturangaben. Gefördert von: Deutsche Forschungsgemeinschaft, Carl-Zeiss-Stiftung.
Abstract (Originalsprache)
We introduce surrogate functionals: machine-learned energy functionals for orbital-free density functional theory (OF-DFT) which are defined not by universal fidelity to a physical reference, but merely by the requirement that density optimization with a fixed procedure yields the true ground-state density. Helpfully, training surrogate functionals requires only ground-state densities as labels, no energies or gradients away from the ground state. We here propose a gradient-descent-improvement loss that guarantees exponential convergence of the density to the ground state, and combine it with an adaptive sampling scheme that concentrates learning around the optimization trajectories actually visited during inference. On the QM9 and QMugs benchmarks, surrogate functionals achieve density errors competitive with or improving upon the state of the art for fully supervised machine-learned OF-DFT, while eliminating the need for the O(N3) orthonormalization step required by prior work, yielding improved runtime scaling for larger systems.
Autor:innen und Institutionen
- Interdisciplinary Center for Scientific Computing (IWR), Heidelberg University, Heidelberg, Germany.
ORCID-Profile: Roman Remme, Fred A Hamprecht
Förderung
- Deutsche Forschungsgemeinschaft
- Carl-Zeiss-Stiftung
Zitieren
APA
Remme, R., & Hamprecht, F. A. (2026). Surrogate functionals for machine-learned orbital-free density functional theory. The Journal of Chemical Physics, 165(3). https://doi.org/10.1063/5.0336653
Vancouver
Remme R, Hamprecht FA. Surrogate functionals for machine-learned orbital-free density functional theory. The Journal of Chemical Physics. 2026;165(3). doi:10.1063/5.0336653
BibTeX
@article{Remme2026,
author = {Remme, Roman and Hamprecht, Fred A},
title = {Surrogate functionals for machine-learned orbital-free density functional theory},
journal = {The Journal of Chemical Physics},
year = {2026},
volume = {165},
number = {3},
doi = {10.1063/5.0336653}
}
EUVIMED