Local pseudopotentials accelerate neural network wavefunction simulations

2026/07/13
Chemistry · Physics and Astronomy · #Advanced Chemical Physics Studies #Advanced Physical and Chemical Molecular Interactions #Artificial neural network #Core (optical fiber) #Electron #Monte Carlo method #Quantum #Quantum Monte Carlo #Quantum and electron transport phenomena #Wave function

paper · doi:10.1038/s43588-026-01027-4

published in Nature Computational Science (Nature Portfolio)

openalex publication_date 2026/07/13 · openalex created_date 2026/07/14 · openalex updated_date 2026/07/23

Abstract

Nature Computational Science, Published online: 13 July 2026; doi:10.1038/s43588-026-01027-4 Local pseudopotentials are shown to remove the cost of core electrons from neural network-based quantum Monte Carlo simulations while avoiding the expensive non-local terms of conventional semi-local pseudopotentials. This strategy increases the computational efficiency and preserves accuracy, enabling studies on previously inaccessible strongly correlated iron–sulfur clusters.

Related

• Empowering Neural Network-based Quantum Monte Carlo with Local Pseudopotentials
• Ab-Initio Solution of the Many-Electron Schrödinger Equation with Deep Neural Networks
• Accurate Ab-initio Neural-network Solutions to Large-Scale Electronic Structure Problems
• Neural Quantum States and Peaked Molecular Wave Functions: Curse or Blessing?
• Quantum Simulation of Realistic Materials in First Quantization Using Non-local Pseudopotentials