mlsanityData to 14 Oct 2026

Long-range electrostatics for machine learning interatomic potentials is easier than we thought

arXiv 2512.18029 ↗First posted 19 Dec 2025physics.comp-ph cond-mat.mtrl-sci cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

The lack of long-range electrostatics is a key limitation of modern machine learning interatomic potentials (MLIPs), hindering reliable applications to interfaces, charge-transfer reactions, polar and ionic materials, and biomolecules. In this Perspective, we distill two design principles behind the Latent Ewald Summation (LES) framework, which can capture long-range interactions, charges, and electrical response just by learning from standard energy and force training data: (i) use a Coulomb functional form with environment-dependent charges to capture electrostatic interactions, and (ii) avoid explicit training on ambiguous density functional theory (DFT) partial charges. When both principles are satisfied, substantial flexibility remains: essentially any short-range MLIP can be augmented; charge equilibration schemes can be added when desired; dipoles and Born effective charges can be inferred or finetuned; and charge/spin-state embeddings or tensorial targets can be further incorporated. We also discuss current limitations and open challenges. Together, these minimal, physics-guided design rules suggest that incorporating long-range electrostatics into MLIPs is simpler and perhaps more broadly applicable than is commonly assumed.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 26 of its 104 references that are in the map. First posted on arXiv 19 Dec 2025.

Builds on

26 of 104 references.

References · 26

  1. 2018Deeper Insights into Graph Convolutional Networks for Semi-Supervised Learningcited by 1,102
  2. 2021E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentialscited by 708
  3. 2020On the Bottleneck of Graph Neural Networks and its Practical Implicationscited by 631
  4. 2017SchNet: A continuous-filter convolutional neural network for modeling quantum interactionscited by 571
  5. 2022MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fieldscited by 474
  6. 2021Equivariant message passing for the prediction of tensorial properties and molecular spectracited by 441
  7. 2022Learning Local Equivariant Representations for Large-Scale Atomistic Dynamicscited by 256
  8. 2020Machine Learning Force Fieldscited by 207
  9. 2025UMA: A Family of Universal Models for Atomscited by 105
  10. 2021SpookyNet: Learning Force Fields with Electronic Degrees of Freedom and Nonlocal Effectscited by 104
  11. 2022The Design Space of E(3)-Equivariant Atom-Centered Interatomic Potentialscited by 101
  12. 2025Learning Smooth and Expressive Interatomic Potentials for Physical Property Predictioncited by 84
14 more
  1. 2022SPICE, A Dataset of Drug-like Molecules and Peptides for Training Machine Learning Potentialscited by 82
  2. 2024Open Materials 2024 (OMat24) Inorganic Materials Dataset and Modelscited by 52
  3. 2023Ewald-based Long-Range Message Passing for Molecular Graphscited by 46
  4. 2024Cartesian atomic cluster expansion for machine learning interatomic potentialscited by 31
  5. 2024Latent Ewald summation for machine learning of long-range interactionscited by 26
  6. 2020Machine learning of solvent effects on molecular spectra and reactionscited by 14
  7. 2023Physics-inspired Equivariant Descriptors of Non-bonded Interactionscited by 14
  8. 2022Capturing long-range interaction with reciprocal space neural networkcited by 13
  9. 2024A predictive machine learning force field framework for liquid electrolyte developmentcited by 13
  10. 2025A universal augmentation framework for long-range electrostatics in machine learning interatomic potentialscited by 9
  11. 2025Machine learning interatomic potential can infer electrical responsecited by 6
  12. 2025Learning Non-Local Molecular Interactions via Equivariant Local Representations and Charge Equilibrationcited by 5
  13. 2025Equivariant Machine Learning Interatomic Potentials with Global Charge Redistributioncited by 4
  14. 2024Broadening the Scope of Neural Network Potentials through Direct Inclusion of Additional Molecular Attributescited by 3

Built on by

Not cited yet.