Hammett-Inspired Product Baseline for Data-efficient $\Delta$-ML in Chemical Space

TL;DR

This paper introduces a Hammett-inspired product baseline model that improves data efficiency in delta-machine learning for chemical property prediction across diverse systems.

Contribution

It presents a novel, general coarse-graining baseline model based on the Hammett equation, applicable to various chemical properties and systems, enhancing delta-ML performance.

Findings

HIP provides accurate baseline predictions across multiple chemical datasets.

HIP improves data efficiency of delta-ML models compared to domain-specific baselines.

Numerical results show HIP's broad applicability and effectiveness.

Abstract

Data-hungry machine learning methods have become a new standard to efficiently navigate chemical compound space for molecular and materials design and discovery. Due to the severe scarcity and cost of high-quality experimental or synthetic simulated training data, however, data-acquisition costs can be considerable. Relying on reasonably accurate approximate legacy baseline labels with low computational complexity represents one of the most effective strategies to curb data-needs, e.g.~through $\Delta$-, transfer-, or multi-fidelity learning. A surprisingly effective and data-efficient baseline model is presented in the form of a generic coarse-graining Hammett-Inspired Product (HIP) {\em Ansatz}, generalizing the empirical Hammett equation towards arbitrary systems and properties. Numerical evidence for the applicability of HIP includes solvation free energies of molecules, formation energies of quaternary elpasolite crystals, carbon adsorption energies on heterogeneous catalytic surfaces, HOMO-LUMO gaps of metallorganic complexes, activation energies for S$_\text{N}$2 reactions, and catalyst-substrate binding energies in cross-coupling reactions. After calibration on the same training sets, HIP yields an effective baseline for improved $\Delta$-machine learning models with superior data-efficiency when compared to previously introduced specialised domain-specific models.