Teaching a Neural Network to Attach and Detach Electrons from Molecules

12 April 2021, Version 2
This content is a preprint and has not undergone peer review at the time of posting.


Physics-inspired Artificial Intelligence (AI) is at the forefront of methods development in molecular modeling and computational chemistry. In particular, interatomic potentials derived with Machine Learning algorithms such as Deep Neural Networks (DNNs), achieve the accuracy of high-fidelity quantum mechanical (QM) methods in areas traditionally dominated by empirical force fields and allow performing massive simulations. The applicability domain of DNN potentials is usually limited by the type of training data. As such, transferable models are aimed to be extensible in the description of chemical and conformational diversity of organic molecules. However, most DNN potentials, such as the AIMNet model we proposed previously, were parametrized for neutral molecules or closed-shell ions due to architectural limitations. In this work, we extend machine learning framework toward open-shell anions and cations. We introduce AIMNet-NSE (Neural Spin Equilibration) architecture, which being properly trained, could predict atomic and molecular properties for an arbitrary combination of molecular charge and spin multiplicity. This model explores a new dimension of transferability by adding the charge-spin space. The AIMNet-NSE model is capable of reproducing reference QM energies for cations, neutrals, and anions with errors of about 2-3 kcal/mol, compared to the reference QM simulations. The spin-charges have errors ~0.01 electrons for small organic molecules containing nine chemical elements {H, C, N, O, F, Si, P, S and Cl}. The AIMNet-NSE model allows to fully bypass QM calculations and derive the ionization potential, electron affinity, and conceptual Density Functional Theory quantities like electronegativity, hardness, and condensed Fukui functions with a speed up to 104 molecules per second on a single modern GPU. We show that these descriptors, along with learned atomic representations, could be used to model chemical reactivity through an example of regioselectivity in electrophilic aromatic substitution reactions.


Conceptual Density Functional Theory
chemical reactivity
machine learning
Neural Networks (NN)


Comments are not moderated before they are posted, but they can be removed by the site moderators if they are found to be in contravention of our Commenting Policy [opens in a new tab] - please read this policy before you post. Comments should be used for scholarly discussion of the content in question. You can find more information about how to use the commenting feature here [opens in a new tab] .
This site is protected by reCAPTCHA and the Google Privacy Policy [opens in a new tab] and Terms of Service [opens in a new tab] apply.