arXiv cs.AI· Yuhao Chen, Donald G. Truhlar, Xiao He·· 13 小时前AI 评分24
平稳性约束同时改善机器学习交换关联泛函的能量与密度
Stationarity Constraints Improve Both Energies and Densities of Machine-Learned Exchange-Correlation Functionals
AI 导读
针对经验交换关联泛函在固定轨道上拟合却自洽使用、能量变好往往密度变差的问题，研究发现训练后泛函在精确密度 Kohn-Sham 轨道上的 Brillouin 残差可一阶控制自洽密度偏移与拟合-自洽能量差，惩罚该项无需在训练中求解自洽场。用 6 个能量和 11 个 CCSD 密度训练的神经 meta-GGA，在未见分子的原子化能与密度上均超过 SCAN。
正文 · 原文
Abstract:Empirical exchange-correlation functionals are fitted on fixed orbitals but used self-consistently, so better energies often come with worse densities. We show that the Brillouin residual of a trained functional on the Kohn-Sham orbitals of an accurate density controls, to first order, both the self-consistent density shift and the gap between fitted and self-consistent energies. Penalizing it requires no self-consistent-field solution during training. Trained on 6 energies and 11 CCSD densities, a neural meta-GGA surpasses SCAN in both atomization energies and densities of unseen molecules.
| Subjects: | Chemical Physics (physics.chem-ph); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.30022 [physics.chem-ph] |
| (or arXiv:2610.30022v1 [physics.chem-ph] for this version) | |
| https://doi.org/10.48550/arXiv.2610.30022 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yuhao Chen [view email]
[v1]
Wed, 21 Oct 2026 16:49:03 UTC (847 KB)
来源：arXiv cs.AI · arxiv.org