Accurate electronic band structures are essential for predicting material properties, but high-fidelity calculations are computationally demanding. We introduce O-kRystal, a dual-space orbital-learnin...
The paper presents a neural operator that learns the Kohn–Sham map, directly predicting electron density from the Kohn–Sham potential without orbital diagonalization. Using a domain‑invariant SE(3)‑equivariant Fourier neural operator trained on 8,504 molecules and solids, the model achieves quasi‑linear scaling self‑consistent field (SCF) convergence across diverse systems—including organic molecules, insulators, and metals—while reproducing Kohn–Sham DFT accuracy for densities, spectra, and structural observables. This enables large‑scale simulations, such as magnesium dislocation densities with 82,500 valence electrons, on a single GPU.
arXiv:2606. 07836v1 Announce Type: cross Abstract: Many-body GW-Bethe-Salpeter equation calculations are essential for accurate simulations of electronic structure and optical properties in modern low-dimensional nanomaterials.
The paper investigates how small entry‑wise errors in machine‑learned Kohn‑Sham Hamiltonians can still lead to inaccurate eigenvalues and eigenvectors. It shows that existing physics‑informed losses only partially capture the error, and proposes using the cumulative local density of states (cumulative LDOS) as a complete training signal that uniquely determines the Hamiltonian. Fine‑tuning three Hamiltonian networks with this loss, called FermiScope, reduces HOMO‑LUMO gap errors without increasing element‑wise error, and the improvement generalizes to out‑of‑distribution and molecular‑dynamics scenarios.
The paper introduces UniTB, a universal, E(3)-equivariant Hamiltonian framework that blends density‑functional tight‑binding priors with deep learning to predict electronic structures efficiently. UniTB corrects a prior Hamiltonian via a mixture‑of‑experts network and a charge‑equilibration head, ensuring physical downstream properties and suppressing spurious states. Trained on over 29,000 crystals, UniTB achieves sub‑meV accuracy, matches DFT predictions for amorphous Ge‑Sb‑Te, and enables rapid screening of more than 16,000 materials for industrial applications.
The paper presents a multitask learning approach that incorporates orbitally resolved semiempirical charges into machine learning interatomic potentials (MLIPs). By training with inexpensive GFN1‑xTB orbital charges, the model reduces energy mean absolute error by 58% compared to energy‑only baselines and needs five times less data to achieve similar performance. The method also outperforms or matches models using costly DFT charges and preserves inference efficiency, offering a scalable way to build accurate, data‑efficient foundation models for complex chemical systems.
The paper introduces MEHnet-MG, an equivariant neural network that predicts a one‑electron Hamiltonian from a single inexpensive B3LYP/def2‑SVP calculation and uses it to compute a wide range of molecular properties—energy, optical gap, dipole, quadrupole, polarizability, Mulliken charges, and Mayer bond orders—at coupled‑cluster accuracy for nine main‑group elements, including phosphorus, sulfur, and chlorine. Trained on a new CCSD(T) dataset, the model reduces property errors by factors ranging from 3.8 to 230 compared to various DFT methods while adding only ~25 ms per molecule. Importantly, by deriving properties from a predicted Hamiltonian rather than pooling atomic features, the architecture naturally incorporates correct size‑scaling, enabling accurate extrapolation to large π‑conjugated systems (up to 58 atoms) where traditional pooling‑based models fail.
arXiv:2608. 14875v1 Announce Type: cross Abstract: Machine-learned exchange-correlation functionals correct band gaps at near-semilocal cost, while density-functional tight binding reaches the $10^3$-$10^6$-atom regime; combining them assumes that a better parent yields a better parameterization, but we show it does not.
The study shows that electronic structure datasets used for machine learning contain significant redundancy due to low intrinsic dimensionality. Random pruning and a coverage-based strategy can reduce dataset size by up to two orders of magnitude while preserving chemical accuracy and generalizability, cutting training time by a factor of three or more. The essential information lies on a low‑dimensional, non‑linear manifold, suggesting that large datasets may be largely overlapping and that minimal, representative datasets could suffice for accurate predictions.
The paper introduces a machine‑learning approach that incorporates the ground‑state electron density as input while enforcing it to be the minimum of a surrogate functional derived from orbital‑free density functional theory. This method outperforms state‑of‑the‑art models on six of seven extensive properties in the QM9 dataset and shows better generalization to larger molecules in the QM40 dataset. Models that use the density without enforcing the surrogate functional constraints perform worse on all extensive properties of both datasets.
arXiv:2607. 25597v1 Announce Type: new Abstract: The reactivity of lithium-metal electrolytes arises from the interplay of molecular functional groups, Li$^+$ solvation, and salt-anion participation.
The paper introduces OrbFlow, an SE(3)-equivariant generative model that predicts Gaussian-type orbital coefficients for electron density using flow matching. By training with a two-phase trajectory curriculum, OrbFlow improves accuracy on QM9 and MD benchmarks, reducing density errors significantly compared to previous models. The predicted densities also accelerate self-consistent field procedures, cutting iterations by up to 68% and accurately reproducing dipole and quadrupole moments without additional SCF calculations.