Learning charges and long-range interactions from energies and forces

• URL: http://arxiv.org/abs/2412.15455v1
• Date: Thu, 19 Dec 2024 23:24:44 GMT
• Title: Learning charges and long-range interactions from energies and forces
• Authors: Dongjin Kim, Daniel S. King, Peichen Zhong, Bingqing Cheng,
• Abstract summary: We introduce the Latent Ewald Summation method, which captures long-range electrostatics without explicitly learning atomic charges or charge equilibration. We benchmark LES on diverse and challenging systems, including charged molecules, ionic liquid, electrolyte solution, polar dipeptides, surface adsorber, electrolyte/solid interfaces, and solid-solid interfaces. Our results show that LES can effectively infer physical partial charges, dipole and quadrupole moments, as well as achieve better accuracy compared to methods that explicitly learn charges.
• Score: 3.502816712907136
• License:
• Abstract: Accurate modeling of long-range forces is critical in atomistic simulations, as they play a central role in determining the properties of materials and chemical systems. However, standard machine learning interatomic potentials (MLIPs) often rely on short-range approximations, limiting their applicability to systems with significant electrostatics and dispersion forces. We recently introduced the Latent Ewald Summation (LES) method, which captures long-range electrostatics without explicitly learning atomic charges or charge equilibration. Extending LES, we incorporate the ability to learn physical partial charges, encode charge states, and the option to impose charge neutrality constraints. We benchmark LES on diverse and challenging systems, including charged molecules, ionic liquid, electrolyte solution, polar dipeptides, surface adsorption, electrolyte/solid interfaces, and solid-solid interfaces. Our results show that LES can effectively infer physical partial charges, dipole and quadrupole moments, as well as achieve better accuracy compared to methods that explicitly learn charges. LES thus provides an efficient, interpretable, and generalizable MLIP framework for simulating complex systems with intricate charge transfer and long-range

Related papers

• Universal Machine Learning Interatomic Potentials are Ready for Solid Ion Conductors [4.7376603774180515]
  Universal machine learning interatomic potentials (uMLIPs) offer a promising solution with their efficiency and near-DFT-level accuracy. This study systematically evaluates six advanced uMLIP models in terms of energy, forces, thermodynamic properties, elastic moduli, and lithium-ion diffusion behavior. The results show that MatterSim outperforms others in nearly all metrics, particularly in complex material systems.
  arXiv  Detail & Related papers  (2025-02-14T07:55:53Z)
• Predicting ionic conductivity in solids from the machine-learned potential energy landscape [68.25662704255433]
  Superionic materials are essential for advancing solid-state batteries, which offer improved energy density and safety. Conventional computational methods for identifying such materials are resource-intensive and not easily scalable. We propose an approach for the quick and reliable evaluation of ionic conductivity through the analysis of a universal interatomic potential.
  arXiv  Detail & Related papers  (2024-11-11T09:01:36Z)
• Latent Ewald summation for machine learning of long-range interactions [0.0]
  We introduce a straightforward and efficient method to account for long-range interactions by learning a latent variable from local atomic descriptors. We demonstrate that in systems including charged and polar molecular dimers, bulk water, and water-vapor interface, standard short-ranged MLIPs can lead to unphysical predictions. The long-range models effectively eliminate these artifacts, with only about twice the computational cost of short-range MLIPs.
  arXiv  Detail & Related papers  (2024-08-27T16:03:18Z)
• Neural Pfaffians: Solving Many Many-Electron Schrödinger Equations [58.130170155147205]
  Neural wave functions accomplished unprecedented accuracies in approximating the ground state of many-electron systems, though at a high computational cost. Recent works proposed amortizing the cost by learning generalized wave functions across different structures and compounds instead of solving each problem independently. This work tackles the problem by defining overparametrized, fully learnable neural wave functions suitable for generalization across molecules.
  arXiv  Detail & Related papers  (2024-05-23T16:30:51Z)
• Modeling Non-Covalent Interatomic Interactions on a Photonic Quantum Computer [50.24983453990065]
  We show that the cQDO model lends itself naturally to simulation on a photonic quantum computer. We calculate the binding energy curve of diatomic systems by leveraging Xanadu's Strawberry Fields photonics library. Remarkably, we find that two coupled bosonic QDOs exhibit a stable bond.
  arXiv  Detail & Related papers  (2023-06-14T14:44:12Z)
• Driving Force and Nonequilibrium Vibronic Dynamics in Charge Separation of Strongly Bound Electron-Hole Pairs [59.94347858883343]
  We study the dynamics of charge separation in one, two and three-dimensional donor-acceptor networks. This allows us to identify the precise conditions in which underdamped vibrational motion induces efficient long-range charge separation.
  arXiv  Detail & Related papers  (2022-05-11T17:51:21Z)
• Relativistic aspects of orbital and magnetic anisotropies in the chemical bonding and structure of lanthanide molecules [60.17174832243075]
  We study the electronic and ro-vibrational states of heavy homonuclear lanthanide Er2 and Tm2 molecules by applying state-of-the-art relativistic methods. We were able to obtain reliable spin-orbit and correlation-induced splittings between the 91 Er2 and 36 Tm2 electronic potentials dissociating to two ground-state atoms.
  arXiv  Detail & Related papers  (2021-07-06T15:34:00Z)
• BIGDML: Towards Exact Machine Learning Force Fields for Materials [55.944221055171276]
  Machine-learning force fields (MLFF) should be accurate, computationally and data efficient, and applicable to molecules, materials, and interfaces thereof. Here, we introduce the Bravais-Inspired Gradient-Domain Machine Learning approach and demonstrate its ability to construct reliable force fields using a training set with just 10-200 atoms.
  arXiv  Detail & Related papers  (2021-06-08T10:14:57Z)
• Many Electrons and the Photon Field -- The many-body structure of nonrelativistic quantum electrodynamics [0.0]
  We show how to turn electronic-structure methods into polaritonic-structure methods that are accurate from the weak to the strong-coupling regime. We discuss how to adopt standard algorithms of electronic-structure methods to adhere to the new hybrid Fermi-Bose statistics.
  arXiv  Detail & Related papers  (2021-02-23T11:00:06Z)
• Multi-scale approach for the prediction of atomic scale properties [0.0]
  Locality underlies machine-learning schemes that predict quantum mechanical observables. One of the main shortcomings of these approaches is their inability to capture physical effects. We show how to build a multi-scale scheme that combines in the same framework local and non-local information.
  arXiv  Detail & Related papers  (2020-08-27T14:01:00Z)
• The Vacua of Dipolar Cavity Quantum Electrodynamics [0.0]
  We show how strong and long-range vacuum fluctuations modify the states of dipolar matter and induce novel phases with unusual properties. These general mechanisms can be important for potential applications, ranging from cavity-assisted chemistry to quantum technologies based on ultrastrongly coupled circuit QED systems.
  arXiv  Detail & Related papers  (2020-04-28T18:00:04Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.