Code and data for "Machine-learning-boosted ab-initio study of the thermal conductivity of Janus PtSTe van der Waals heterostructures"

<h1>Code and data for <em>Machine-learning-boosted ab-initio study of the thermal conductivity of Janus PtSTe van der Waals heterostructures</em></h1> <p>&nbsp;</p> <h2>Contents:</h2> <ul> <li><strong>neuralil.tar.xz</strong>: version used in the manuscript of the force-field code described in the articles&nbsp; <a href="https://doi.org/10.1021/acs.jcim.1c01380">A Differentiable Neural-Network Force Field for Ionic Liquids</a> and <a href="https://doi.org/10.1063/5.0146905">Deep ensembles vs committees for uncertainty estimation in neural-network force fields: Comparison and application to active learning</a>. General-purpose releases can be found <a href="https://github.com/Madsen-s-research-group/neuralil-public-releases">here</a>.</li> <li><strong>DFT_data.tar.xz</strong>: first-principles data created for training and validating the force field, stored as <a href="https://wiki.fysik.dtu.dk/ase/ase/db/db.html">ASE databases</a> in JSON format.</li> <li><strong>model_params_plain_ensemble_DEEP_413E12A9.pkl</strong>: saved parameters of the fully trained force field.</li> <li><strong>0001-Use-equipartition-occupancies.patch</strong>: patch for <a href="https://phonopy.github.io/phono3py">Phono3py</a> to use classical (equipartition) occupations instead of Bose-Einstein values.</li> </ul>