M-OFDFT: Overcoming the Barrier of Orbital-Free Density Functional Theory for Molecular Systems Using Deep Learning

• DOI: 10.48550/arxiv.2309.16578
• Publication Date: 2023-01-01
• Authors: He Zhang; Siyuan Liu; Jiacheng You; Chang Liu; Shuxin Zheng; Ziheng Lu; Tong Wang; Nanning Zheng; Bin Shao
• Keywords: Basis (linear algebra); Representation; Quantum Chemistry

Orbital-free density functional theory (OFDFT) is a quantum chemistry formulation that has lower cost scaling than the prevailing Kohn-Sham DFT, which increasingly desired for contemporary molecular research. However, its accuracy limited by kinetic energy functional, notoriously hard to approximate non-periodic systems. Here we propose M-OFDFT, an OFDFT approach capable of solving systems using deep learning model. We build essential non-locality into model, made affordable concise representation as expansion coefficients under atomic basis. With techniques address unconventional challenges therein, M-OFDFT achieves comparable with DFT on wide range molecules untouched before. More attractively, extrapolates well much larger those seen in training, unleashes appealing studying large including proteins, representing advancement accuracy-efficiency trade-off frontier chemistry.