Machine learning evaluation of structural descriptors for supercooled water

TL;DR

This study systematically evaluates 16 structural descriptors of supercooled water using neural networks and explainable AI, providing a data-driven benchmark for understanding local structural features.

Contribution

It introduces a neural-network framework to objectively compare and interpret various structural descriptors of supercooled water.

Findings

Neural network effectively classifies temperature based on structural descriptors.

Explainable AI identifies key structural features influencing model predictions.

Benchmarking reveals strengths and limitations of existing descriptors.

Abstract

The anomalous behavior of liquid water is widely associated with a liquid-liquid phase transition between high- and low-density states in the supercooled regime. At the microscopic level, tetrahedral hydrogen-bond networks govern these properties, motivating structural descriptors that characterize local molecular environments. These structural descriptors quantify features such as tetrahedral order, local density, and the separation between the first and second coordination shells; however, they have largely been proposed independently, with limited systematic comparison. Here we evaluate 16 previously proposed descriptors using a neural-network-based temperature classification framework, enabling an objective assessment of their ability to distinguish temperature-dependent structural changes in supercooled water. We further apply an explainable artificial intelligence method that identifies the structural features responsible for the model predictions. This approach reveals how different descriptors encode local structural information and establishes a data-driven framework for benchmarking structural descriptors in liquid water.