2026-10-13T20:21:18Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33820543 2026-09-15T19:55:00Z portal_1197 item_type_3 month_year_09_2026

Data for Combining phonon accuracy with high transferability in Gaussian approximation potential models Janine George (1680829) Geoffery Hautier (24954067) Albert P. Bartók (8873483) Gábor Csányi (2475193) Volker L. Deringer (1889818) Gaussian processes Phonons Atomic structure Vibrational spectra Configuration space Materials -- Mathematical models Materials science -- Computer simulation Migrated from ePrints Machine learning driven interatomic potentials, including Gaussian approximation potential (GAP) models, are emerging tools for atomistic simulations. Here, we address the methodological question of how one can fit GAP models that accurately predict vibrational properties in specific regions of configuration space while retaining flexibility and transferability to others. We use an adaptive regularization of the GAP fit that scales with the absolute force magnitude on any given atom, thereby exploring the Bayesian interpretation of GAP regularization as an "expected error" and its impact on the prediction of physical properties for a material of interest. The approach enables excellent predictions of phonon modes (to within 0.1 THz-0.2 THz) for structurally diverse silicon allotropes, and it can be coupled with existing fitting databases for high transferability across different regions of configuration space, which we demonstrate for liquid and amorphous silicon. These findings and workflows are expected to be useful for GAP-driven materials modeling more generally.<br><br>Data record consists of a zip archive, organised according to figure, containing the raw data in xml format, as well as an accompanying readme file. All models are defined by the main parameter file "gp_iter6C.xml" and the associated file "gp_iter6C.xml.sparseX.GAP_00001", where "GAP_0000" is a placeholder for the unique identifier of the potential (also given in the XML header), and the trailing "1" indicates that only one set of descriptor (here, SOAP) parameters is given. The directories in this dataset follow the figures in the publication for which the respective potentials have been first used. 2026-09-15T19:55:00Z Dataset Dataset 10.82444/warw.33820543.v1 https://figshare.com/articles/dataset/Data_for_Combining_phonon_accuracy_with_high_transferability_in_Gaussian_approximation_potential_models/33820543 CC BY 4.0