English

Neural density functionals: Local learning and pair-correlation matching

Soft Condensed Matter 2024-09-16 v3 Statistical Mechanics

Abstract

Recently Dijkman et al. (arXiv:2403.15007) proposed training classical neural density functionals via bulk pair-correlation matching. We show their method to be an efficient regularizer for neural functionals based on local learning of inhomogeneous one-body direct correlations [Samm\"uller et al., Proc. Natl. Acad. Sci. 120, e2312484120 (2023), 10.1073/pnas.2312484120]. While Dijkman et al. demonstrated pair-correlation matching of a global neural free energy functional, we argue in favor of local one-body learning for flexible neural modelling of the full Mermin-Evans density functional map. Using spatial localization gives access to accurate neural free energy functionals, including convolutional neural networks, that transcend the training box.

Keywords

Cite

@article{arxiv.2406.03327,
  title  = {Neural density functionals: Local learning and pair-correlation matching},
  author = {Florian Sammüller and Matthias Schmidt},
  journal= {arXiv preprint arXiv:2406.03327},
  year   = {2024}
}

Comments

6 pages, 2 figures + supplementary material (7 pages, 6 figures)

R2 v1 2026-06-28T16:54:38.782Z