Complete Neural Electronic Initialization Accelerates Materials DFT
This research provides a new approach to accelerate a fundamental computational method used in materials science and chemistry, which could lead to faster and more efficient simulations.
This paper introduces a new method called Complete Neural Electronic Initialization to speed up plane-wave density functional theory (DFT) calculations for materials.
It addresses two missing components in prior work: augmentation occupancies and spin initialization.The method uses a model called AugNet, the first general equivariant model for PAW augmentation occupancies, and a spin density model.
When combined with existing valence density models, it forms a fully reference-free electronic initializer that reduces end-to-end DFT wall time by up to 25% on unseen structures while preserving converged energies.
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 \textit{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.
This report is categorized under AITOP's "Papers" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Key entities and technical topics under continuous tracking include density functional theory, DFT acceleration, machine learning models, augmentation occupancies.
Source attribution: curated from arXiv cs.LG (@Felix Ærtebjerg, Jonas Elsborg, Arghya Bhowmik).