Inferring the dynamics of glass-forming liquids from static structure across thermal states

TL;DR

This paper introduces T-BOTAN, a graph neural network framework that predicts dynamic heterogeneity and thermodynamic properties of glass-forming liquids from static structures across different temperatures, demonstrating strong generalization.

Contribution

The study presents a novel GNN-based method that accurately infers dynamics and temperature from static configurations, extending predictions beyond trained temperatures.

Findings

T-BOTAN accurately predicts dynamics at untrained temperatures.

Static structures encode thermodynamic and dynamic information.

The method generalizes across different thermal states.

Abstract

In this study, we demonstrate the generalizability of graph neural networks in predicting the dynamic heterogeneity of model glass-forming liquids across different temperatures. While previous approaches have often been limited to making predictions at the specific temperatures used during training, we find that our proposed framework - T-BOTAN - enables interpolation to temperatures not included in the training set. We show that the dynamical behavior, the associated four-point correlations, and even the macroscopic temperature can be estimated with sufficient accuracy solely from static particle configurations at untrained temperatures. These results suggest that static configurations encode not only local structural features driving dynamic heterogeneity but also fundamental thermodynamic information.