Electronic Descriptors for Supervised Spectroscopic Predictions

Spectroscopic properties of molecules holds great importance for the description of the molecular response under the effect of an UV/Vis electromagnetic radiation. Computationally expensive ab initio (e.g. MultiConfigurational SCF, Coupled Cluster) or TDDFT methods are commonly used by the quantum chemistry community to compute these properties. In this work, we propose a (supervised) Machine Learning approach to model the absorption spectra of organic molecules. Several supervised ML methods have been tested such as Kernel Ridge Regression (KRR), Multiperceptron Neural Networs (MLP) and Convolutional Neural Networks. The use of only geometrical descriptors (e.g. Coulomb Matrix) proved to be insufficient for an accurate training. Inspired on the TDDFT theory, we propose to use a set of electronic descriptors obtained from low-cost DFT methods: orbital energy differences, transition dipole moment between occupied and unoccupied Kohn-Sham orbitals and charge-transfer character of mono-excitations. We demonstrate that with this electronic descriptors and the use of Neural Networks we can predict not only a density of excited states, but also getting very good estimation of the absorption spectrum and charge-transfer character of the electronic excited states, reaching results close to the chemical accuracy (~2 kcal/mol or ~0.1eV).