Understanding molecular representations in machine learning: The role of uniqueness and target similarity

TL;DR

This paper introduces a hierarchy of molecular representations based on quantum mechanics principles, demonstrating that including higher-order terms improves the accuracy of machine learning models for predicting various molecular properties.

Contribution

It proposes a systematic approach to control target similarity in molecular representations using many-body expansions, enhancing ML prediction accuracy for molecular properties.

Findings

Higher-order representation terms improve predictive accuracy.

BAML models outperform existing methods in speed and accuracy.

Systematic inclusion of expansion terms benefits diverse molecular properties.

Abstract

The predictive accuracy of Machine Learning (ML) models of molecular properties depends on the choice of the molecular representation. Based on the postulates of quantum mechanics, we introduce a hierarchy of representations which meet uniqueness and target similarity criteria. To systematically control target similarity, we rely on interatomic many body expansions, as implemented in universal force-fields, including Bonding, Angular, and higher order terms (BA). Addition of higher order contributions systematically increases similarity to the true potential energy and predictive accuracy of the resulting ML models. We report numerical evidence for the performance of BAML models trained on molecular properties pre-calculated at electron-correlated and density functional theory level of theory for thousands of small organic molecules. Properties studied include enthalpies and free energies of atomization, heatcapacity, zero-point vibrational energies, dipole-moment, polarizability, HOMO/LUMO energies and gap, ionization potential, electron affinity, and electronic excitations. After training, BAML predicts energies or electronic properties of out-of-sample molecules with unprecedented accuracy and speed.