# Machine Learning Unifies the Modelling of Materials and Molecules

## TL;DR

This paper introduces a machine learning framework that unifies the modeling of materials and molecules, accurately predicting atomic properties and stability across different chemical systems using a local environment description and Bayesian learning.

## Contribution

It presents a novel, universal machine learning approach that captures quantum effects and predicts stability and activity with high accuracy, advancing atomistic modeling.

## Key findings

- Predicts silicon surface reconstructions with quantum accuracy
- Accurately assesses molecular stability across classes
- Distinguishes active/inactive protein ligands with >99% reliability

## Abstract

Determining the stability of molecules and condensed phases is the cornerstone of atomistic modelling, underpinning our understanding of chemical and materials properties and transformations. Here we show that a machine learning model, based on a local description of chemical environments and Bayesian statistical learning, provides a unified framework to predict atomic-scale properties. It captures the quantum mechanical effects governing the complex surface reconstructions of silicon, predicts the stability of different classes of molecules with chemical accuracy, and distinguishes active and inactive protein ligands with more than 99% reliability. The universality and the systematic nature of our framework provides new insight into the potential energy surface of materials and molecules.