发表机构
Quemix Inc.; The University of Tokyo; JSR Corporation; National Institutes for Quantum Science and Technology (QST); RIKEN Center for Computational Science (R-CCS)（Quemix公司; 东京大学; JSR公司; 国立量子科学技术研究院; 理化学研究所计算科学中心）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
提出一种基于原子分解和图神经网络的偶极矩预测模型，可准确预测复杂共轭体系的偶极矩与太赫兹介电函数，推理速度比玻恩有效电荷方案快约3倍。
AI 中文摘要
我们提出了一种预测分子体系偶极矩的方法，该方法利用密度泛函理论计算得到的波函数，将总偶极矩系统地分解为原子贡献之和，然后使用图神经网络根据输入的原子结构来预测这一有效的原子偶极矩。结果表明，即使对于先前基于键的模型[Phys. Rev. B 110, 165159]失效的复杂共轭体系，该方法也能准确预测偶极矩和介电函数。我们模型的推理成本随原子数量线性增长，并且比基于玻恩有效电荷的方案快约3倍，但在太赫兹范围内介电函数的精度相似或更优。
英文摘要
We introduce a method for predicting the dipole moments of molecular systems by systematically decomposing the total dipole moment into the sum of atomic contributions using the wavefunction from density-functional-theory calculations, and then use graph neutral networks to predict this effective atomic dipole moment from input atomic structure. It is demonstrated that the dipole moments and dielectric function can be accurately predicted even for complicated conjugated systems where the previous bond-based model [Phys. Rev. B 110, 165159] fails. The inference cost of our model scales linearly with the number of atoms and is about 3 times faster compared with Born-effective-charge-based schemes, but shows similar or better accuracy for dielectric function at terahertz range.
