Decomposing Chemical Space: Applications to the Machine Learning of Atomic Energies

TL;DR

This paper compares different atomic decomposition schemes on the QM7 dataset to understand how they reveal trends in atomic contributions to molecular energies and their implications for machine learning models.

Contribution

It introduces and evaluates various atomic decomposition methods, highlighting their differences and impacts on machine learning of atomic energies.

Findings

Decomposition schemes reveal diverse atomic contribution patterns.

Some schemes show unphysical basis set dependencies.

Insights inform the design of atomic-energy-based neural network models.

Abstract

We apply a number of atomic decomposition schemes across the standard QM7 dataset -- a small model set of organic molecules at equilibrium geometry -- to inspect the possible emergence of trends among contributions to atomization energies from distinct elements embedded within molecules. Specifically, a recent decomposition scheme of ours based on spatially localized molecular orbitals is compared to alternatives that instead partition molecular energies on account of which nuclei individual atomic orbitals are centred on. We find these partitioning schemes to expose the composition of chemical compound space in very dissimilar ways in terms of the grouping, binning, and heterogeneity of discrete atomic contributions, e.g., those associated with hydrogens bonded to different heavy atoms. Furthermore, unphysical dependencies on the one-electron basis set are found for some, but not all of these schemes. The relevance and importance of these compositional factors for training tailored neural network models based on atomic energies are next assessed. We identify both limitations and possible advantages with respect to contemporary machine learning models and discuss the design of potential counterparts based on atoms and the intrinsic energies of these as the principal decomposition units.