Local Pseudopotential Unlocks the True Potential of Neural Network-based Quantum Monte Carlo
Abstract
Neural Network-based Quantum Monte Carlo (NNQMC), an emerging method for solving many-body quantum systems with high accuracy, has been limitedly applied to small systems due to demanding computation requirements. In this work, we introduce an approach based on local pseudopotentials to break through such limitation, significantly improving the computational efficiency and scalability of NNQMC. The incorporation of local pseudopotentials not only reduces the number of electrons treated in neural network but also achieves better accuracy than all electron NNQMC calculations for complex systems. This counterintuitive outcome is made possible by the distinctive characteristics inherent to NNQMC. Our approach enables the reliable treatment of large and challenging systems, such as iron-sulfur clusters with as many as 268 total electrons, which were previously beyond reach for NNQMC methods. Overall, our findings demonstrate that the synergy between NNQMC and local pseudopotentials substantially expands the scope of accurate ab initio calculations, pushing the frontiers of quantum chemistry and computational physics.
Keywords
Cite
@article{arxiv.2505.19909,
title = {Local Pseudopotential Unlocks the True Potential of Neural Network-based Quantum Monte Carlo},
author = {Weizhong Fu and Ryunosuke Fujimaru and Ruichen Li and Yuzhi Liu and Xuelan Wen and Xiang Li and Kenta Hongo and Liwei Wang and Tom Ichibha and Ryo Maezono and Ji Chen and Weiluo Ren},
journal= {arXiv preprint arXiv:2505.19909},
year = {2025}
}