Electron-Density-Informed Molecular Property Prediction Using Surrogate Functionals

Abstract

Machine learning methods are increasingly used to predict chemical properties directly from the molecular structure to bypass the often prohibitive cost of quantum mechanical calculations. Recent work has shown that including the ground-state electron density as input to the machine learning model can improve accuracy in this setting. This work employs the recent surrogate functional framework originally developed for orbital-free density functional theory and extends it to predicting molecular properties. We show that passing the ground-state density as input to the model, while simultaneously enforcing it to be the minimum of the surrogate functional, surpasses state-of-the-art performance for six out of seven extensive properties of the QM9 dataset. Additionally, the approach shows improved generalization to larger molecules taken from the QM40 dataset. On all extensive properties of both datasets, our approach outperforms models that are informed by the ground-state density but do not enforce the surrogate functional constraints.

Orbital-Free Surrogate Functionals Yield Transferable Interatomic Potentials and Electron Densities
Simon Wagner, Marc K. Ickler, Manuel V. Klockow +2
arXiv18 Sep 2026PreprintarXiv licenceFull textPhysical chemistry Computational physicsarXiv
Physics-Informed Deep Learning Surrogate for Mechanical-Response Prediction and Failure Assessment of High-Performance Polymer Composites
Md Ehtesam Haque, Ramasamy Naidu, Md Arifuzzaman +3
Polymers28 Oct 2026Peer-reviewedCC BY 4.0Materials & polymersCrossref
DeMiR: Density-Matrix-Informed Representations for Molecular Property Prediction
Donghyeon Kim, Dongsup Kim
ChemRxiv5 Oct 2026PreprintCC BYFull textGeneral chemistryChemRxiv
Accelerating Phase Prediction via CALPHAD-Informed Machine Learning and Data Augmentation
Caroline Binde Stoco, Guillaume Deffrennes, Yannick Champion +1
Computational Materials Science4 Jun 2026Peer-reviewedLicence Ouverte 2.0Full text5 citedMaterials & polymersHALOpenAlex
Electrochemical Proton-Coupled Electron Transfer Free Energy Surfaces from Ab Initio Molecular Dynamics
Mahmudul Hasan, Robert E. Warburton
ChemRxiv2 Sep 2026PreprintLicence not statedGeneral chemistryChemRxiv

Cite

BibTeX RIS

BibTeX

@misc{klockow2026electrondensityinformed,
  title = {{Electron-Density-Informed Molecular Property Prediction Using Surrogate Functionals}},
  author = {Manuel V. Klockow and Simon Wagner and Roman Remme and Dominic Plein and Fred A. Hamprecht},
  howpublished = {arXiv},
  year = {2026},
  month = oct,
  doi = {10.48550/arxiv.2610.13649},
  eprint = {2610.13649},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2610.13649},
}

RIS

TY  - UNPB
TI  - Electron-Density-Informed Molecular Property Prediction Using Surrogate Functionals
AU  - Manuel V. Klockow
AU  - Simon Wagner
AU  - Roman Remme
AU  - Dominic Plein
AU  - Fred A. Hamprecht
PY  - 2026
DA  - 2026/10/09
PB  - arXiv
DO  - 10.48550/arxiv.2610.13649
UR  - https://arxiv.org/abs/2610.13649
AB  - Machine learning methods are increasingly used to predict chemical properties directly from the molecular structure to bypass the often prohibitive cost of quantum mechanical calculations. Recent work has shown that including the ground-state electron density as input to the machine learning model can improve accuracy in this setting. This work employs the recent surrogate functional framework originally developed for orbital-free density functional theory and extends it to predicting molecular properties. We show that passing the ground-state density as input to the model, while simultaneously enforcing it to be the minimum of the surrogate functional, surpasses state-of-the-art performance for six out of seven extensive properties of the QM9 dataset. Additionally, the approach shows improved generalization to larger molecules taken from the QM40 dataset. On all extensive properties of both datasets, our approach outperforms models that are informed by the ground-state density but do not enforce the surrogate functional constraints.
ER  -

Abstract

Related

Cite