Solid Harmonic Wavelet Scattering for Predictions of Molecule Properties

TL;DR

This paper introduces a machine learning approach using solid harmonic wavelet scattering coefficients derived from surrogate electronic densities to predict molecular properties with high accuracy and interpretability.

Contribution

It proposes a novel invariant scattering coefficient method inspired by density functional theory for molecular property prediction.

Findings

Achieves near DFT precision in property predictions.

Performs well with limited training data.

Provides interpretable results.

Abstract

We present a machine learning algorithm for the prediction of molecule properties inspired by ideas from density functional theory. Using Gaussian-type orbital functions, we create surrogate electronic densities of the molecule from which we compute invariant "solid harmonic scattering coefficients" that account for different types of interactions at different scales. Multi-linear regressions of various physical properties of molecules are computed from these invariant coefficients. Numerical experiments show that these regressions have near state of the art performance, even with relatively few training examples. Predictions over small sets of scattering coefficients can reach a DFT precision while being interpretable.