Neural operators learn Kohn-Sham map for faster DFT calculations

Researchers have developed a novel approach to density functional theory (DFT) by using neural operators to learn the Kohn-Sham map, which bypasses the computationally intensive orbital diagonalization step. This method, trained on a large dataset of molecules and solids, can predict electron density and non-interacting kinetic energy, enabling stable quasi-linear scaling self-consistent field (SCF) calculations. The trained model demonstrates generalization to out-of-distribution systems and accurately reproduces densities and electronic spectra at Kohn-Sham DFT accuracy, allowing for the convergence of SCFs for systems with tens of thousands of electrons on a single GPU. AI

IMPACT Enables faster and more scalable electronic-structure simulations, potentially accelerating materials science and drug discovery.

RANK_REASON Academic paper detailing a new method for density functional theory calculations. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• density functional theory
• Fourier Neural Operator
• Kohn–Sham equations
• SE(3)-Transformers: 3D Roto-Translation Equivariant Attention Networks

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →