Data package accompanying the publication "Improving the prediction of glassy dynamics by pinpointing the local cage" in the Journal of Chemical Physics. This package provides all the relevant information for the analysis on the cage state associated with three different glass formers (binary hard spheres, binary harmonic mixture, and binary Kob-Andersen mixture). In the paper we explore whether a simple linear regression algorithm combined with intelligently chosen structural order parameters can reach the accuracy of the current, most advanced machine learning approaches for predicting dynamic propensity. To do this we introduce a method to pinpoint the cage state of the initial configuration – i.e. the configuration consisting of the average particle positions when particle rearrangement is forbidden. We find that, in comparison to both the initial state and the inherent state, the structure of the cage state is highly predictive of the long-time dynamics of the system. Moreover, by combining the cage state information with the initial state, we are able to predict dynamic propensities with unprecedentedly high accuracy over a broad regime of time scales, including the caging regime.
In each directory there are READ_ME files that describe the content. The main directory contains four directories: the directory [FIGURES], which contains the published graphs, the directory [CODES], which contains the codes the obtain the inherent state, the cage state and the structural parameters as described in the paper, the directory [DATA] which contains all the necessary data to do the analysis described in the paper for the three different glassy systems and the directory [OUTPUT], which contains the linear regression files to train the linear regression models, and the associated learned correlations between the measured and predicted propensities for the three different glassy systems.
