Data for the manuscript "Spatially resolved uncertainties for machine learning potentials"

<p>This repository accompanies the manuscript "Spatially resolved uncertainties for machine learning potentials" by E. Heid, J. Sch&ouml;rghuber, R. Wanzenb&ouml;ck, and G. K. H. Madsen. The following files are available:</p> <ul> <li> <p><code>mc_experiment.ipynb</code> is a Jupyter notebook for the Monte Carlo experiment described in the study (artificial model with only variance as error source).</p> </li> </ul> <ul> <li> <p><code>aggregate_cut_relax.py</code> contains code to cut and relax boxes for the water active learning cycle.</p> </li> <li> <p><code>data_t1x.tar.gz</code> contains reaction pathways for 10,073 reactions from a subset of the Transition1x dataset, split into training, validation and test sets. The training and validation sets contain the indices 1, 2, 9, and 10 from a 10-image nudged-elastic band search (40k datapoints), while the test set contains indices 3-8 (60k datapoints). The test set is ordered according to the reaction and index, i.e. rxn1_index3, rxn1_index4, [...] rxn1_index8, rxn2_index3, [...].</p> </li> <li> <p><code>data_sto.tar.gz</code> contains surface reconstructions of SrTiO3, randomly split into a training and validation set, as well as a test set.</p> </li> <li> <p><code>data_h2o.tar.gz</code> contains:</p> <ul> <li> <p><code>full_db.extxyz</code>: The full dataset of 1.5k structures.</p> </li> <li> <p><code>iter00_train.extxyz</code> and <code>iter00_validation.extxyz</code>: The initial training and validation set for the active learning cycle.</p> </li> <li> <p>the subfolders in the folders <code>random</code>, and <code>uncertain</code>, and <code>atomic</code> contain the training and validation sets for the random and uncertainty-based (local or atomic) active learning loops.</p> </li> </ul> </li> </ul>