Complete Neural Electronic Initialization Accelerates Materials DFT
cond-mat.mtrl-sci, cs.LG, physics.comp-ph
Submitted: 2026-09-18
Updated: 2026-09-27
Comments: 34 pages, 4 figures, 15 tables
Code: https://github.com/aerte/neural_paw_dft
License: http://creativecommons.org/licenses/by/4.0/
The gist: We present the first complete machine learning method for accelerating plane-wave density functional theory (DFT) in materials under the projector augmented wave (PAW) formalism.
Terminology
Abstract
We present the first complete machine learning method for accelerating plane-wave density functional theory (DFT) in materials under the projector augmented wave (PAW) formalism. We formalize seven criteria that a Complete Neural Electronic Initializer must satisfy for practical end-to-end PAW DFT acceleration. Applying these criteria to prior work reveals two missing structure-dependent components, augmentation occupancies and spin initialization, that prevent existing methods from providing complete reference-free initialization. Controlled ablations show that omitting these components can eliminate or reverse the acceleration obtained via models that only predict the smooth valence density. We satisfy these missing requirements by introducing AugNet, the first general equivariant model for PAW augmentation occupancies, and the first general spin density model for materials, which predicts the smooth spin-difference density and spin-difference PAW augmentation occupancies using predicted magnetic moments to constrain the global magnetic state. Combined with existing valence density models, these components satisfy all seven criteria and form a fully reference-free electronic initializer for materials DFT, requiring no electronic quantities from a converged target calculation. Our method reduces end-to-end DFT wall time by up to 25% on unseen structures while preserving converged energies.
Sources
- Approximating with Gaussians
- Transferable SCF-Acceleration through Solver-Aligned Initialization Learning
- Global Plane Waves From Local Gaussians: Periodic Charge Densities in a Blink
- Learning from the electronic structure of molecules across the periodic table
- Gaussian Plane-Wave Neural Operator for Electron Density Estimation
- Machine Learning Hamiltonians are Accurate Energy-Force Predictors
- Efficient E(3)-equivariant framework for universal charge density prediction
- Orb: A Fast, Scalable Neural Network Potential
- UMA: A Family of Universal Models for Atoms
Related papers
- AES-Debye: an Accurate, Efficient, and Scalable Engine for Debye Scattering Calculations
- Cooperative Quantum Optical Effects of Moir'e Exciton Superlattices
- Imaging Surface Magnetization in Altermagnetic MnTe Films
- Accidental accuracy and formal consistency in GW +BSE: Exact benchmarks and regime-dependent error cancellation
- Modifying van der Waals Materials via Cavity Vacuum Fluctuations
- Linear dichroic soft X-ray microscopy of ferroelectric stripe domains in epitaxial K 0.6 Na 0.4 NbO 3