Dresden 2020 – wissenschaftliches Programm

Die DPG-Frühjahrstagung in Dresden musste abgesagt werden! Lesen Sie mehr ...

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

CPP: Fachverband Chemische Physik und Polymerphysik

CPP 107: Emerging Topics in Chemical and Polymer Physics, New Instruments and Methods

CPP 107.2: Vortrag

Freitag, 20. März 2020, 10:00–10:15, ZEU 255

Machine learning the molecular dipole moment with atomic partial charges and atomic dipoles — •Max Veit¹, David Wilkins¹, Yang Yang², Robert DiStasio Jr², and Michele Ceriotti¹ — ¹Laboratory of Computational Science and Modeling, IMX, École Polytechnique Fédérale de Lausanne, 1015 Lausanne, Switzerland — ²Department of Chemistry and Chemical Biology, Cornell University, Ithaca, NY 14853, USA

The gas-phase molecular dipole moment is a central quantity in chemistry. It is essential in predicting molecular infrared and sum-frequency-generation spectra, as well as in describing long-range molecular interactions. Furthermore, it can be extracted directly from expensive, but highly accurate, quantum mechanical calculations, making it an ideal target for machine learning. We choose to represent this quantity with a physically-inspired machine learning model that captures the two distinct physical effects contributing to molecular polarization: Local atomic polarization is captured within the symmetry-adapted Gaussian process regression (SA-GPR) framework, which assigns a dipole moment to each atom, while movement of charge across the entire molecule is captured by assigning a partial (scalar) charge to each atom. These models are fitted together to reproduce the quantum mechanical molecular dipole moment, achieving better results than either model alone. The results show how transparency and physical interpretability can aid not only the understanding of a machine learning model, but allow it to achieve higher accuracy as well, regardless of which physical property is being modelled.