An atom-based machine-learned dipole-moment model and application to conjugated systems
Abstract
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.
Related
- All-electron machine-learned orbital-free density functional for periodic CHNO systems
- Benzylic Oxidation of Pillar[5]Arene: From Homo‐ to Cross‐Conjugated Macrocyclic Systems
- Excited states engineering maximizes singlet generation by triplet fusion in conjugated systems
- Synthesis of Tris(indol‐3‐yl)methanes From Indoles, CO2, and Dimethylphenylsilane With BPh3: Two‐Step Construction of π‐Conjugated Systems
- Decoding Chemical Resonance in π-Conjugated Systems
BibTeX
@misc{zhao2026atombased,
title = {{An atom-based machine-learned dipole-moment model and application to conjugated systems}},
author = {Yuansheng Zhao and Tamio Yamazaki and Ryohei Hosoya and Yu-ichiro Matsushita and Shinji Tsuneyuki},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.39002},
eprint = {2609.39002},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.39002},
}
RIS
TY - UNPB
TI - An atom-based machine-learned dipole-moment model and application to conjugated systems
AU - Yuansheng Zhao
AU - Tamio Yamazaki
AU - Ryohei Hosoya
AU - Yu-ichiro Matsushita
AU - Shinji Tsuneyuki
PY - 2026
DA - 2026/09/30
PB - arXiv
DO - 10.48550/arxiv.2609.39002
UR - https://arxiv.org/abs/2609.39002
AB - 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.
ER -