Machine learning sparse tight-binding parameters for defects

dc.contributorAalto-yliopistofi
dc.contributorAalto Universityen
dc.contributor.authorSchattauer, Christophen_US
dc.contributor.authorTodorović, Milicaen_US
dc.contributor.authorGhosh, Kunalen_US
dc.contributor.authorRinke, Patricken_US
dc.contributor.authorLibisch, Florianen_US
dc.contributor.departmentDepartment of Applied Physicsen
dc.contributor.departmentDepartment of Computer Scienceen
dc.contributor.groupauthorComputational Electronic Structure Theoryen
dc.contributor.organizationTechnische Universität Wienen_US
dc.date.accessioned2022-08-10T08:28:01Z
dc.date.available2022-08-10T08:28:01Z
dc.date.issued2022-05-20en_US
dc.descriptionFunding Information: We acknowledge support from the FWF DACH project I3827-N36, COST action CA18234, the Academy of Finland through projects 316601 and 334532 and the doctoral colleges Solids4Fun W1243-N16 funded by the FWF and TU-D funded by TU Wien. Christoph Schattauer acknowledges support as a recipient of a DOC fellowship of the Austrian Academy of Sciences. Numerical calculations were performed on the Vienna Scientific Clusters VSC3 and VSC4. Funding Information: iPALM imaging was done in collaboration with the Advanced Imaging Center at Janelia Research Campus, a facility jointly supported by the Gordon and Betty Moore Foundation and the Howard Hughes Medical Institute. ChIA-Drop analysis was done by Minji Kim and Michał Denkiewicz. Molecular graphics and analyses performed with UCSF Chimera, developed by the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco, with support from NIH P41-GM103311. This work was carried out within „Three-dimensional Human Genome structure at the population scale: computational algorithm and experimental validation for lymphoblastoid cell lines of selected families from 1000 Genomes Project” project carried out within the TEAM programme of the Foundation for Polish Science co-financed by the European Union under the European Regional Development Fund, co-supported by Polish National Science Centre (2019/35/O/ST6/02484 and 2020/37/B/NZ2/03757) and the grant 1U54DK107967-01 “Nucleome Positioning System for Spatiotemporal Genome Organization and Regulation” within 4D Nucleome NIH program. The work was co-supported by the European Commission as Horizon 2020 Marie Skłodowska-Curie ITN Enhpathy grant “Molecular Basis of Human enhanceropathies”. DP was co-funded by Warsaw University of Technology within the Excellence Initiative: Research University (IDUB) programme; and POB Cybersecurity and data analysis of Warsaw University of Technology within the Excellence Initiative: Research University (IDUB) programme. Computations were performed thanks to the Laboratory of Bioinformatics and Computational Genomics, Faculty of Mathematics and Information Science, Warsaw University of Technology using Artificial Intelligence HPC platform financed by Polish Ministry of Science and Higher Education (decision no. 7054/IA/SP/2020 of 2020-08-28). Publisher Copyright: © 2022, The Author(s).
dc.description.abstractWe employ machine learning to derive tight-binding parametrizations for the electronic structure of defects. We test several machine learning methods that map the atomic and electronic structure of a defect onto a sparse tight-binding parameterization. Since Multi-layer perceptrons (i.e., feed-forward neural networks) perform best we adopt them for our further investigations. We demonstrate the accuracy of our parameterizations for a range of important electronic structure properties such as band structure, local density of states, transport and level spacing simulations for two common defects in single layer graphene. Our machine learning approach achieves results comparable to maximally localized Wannier functions (i.e., DFT accuracy) without prior knowledge about the electronic structure of the defects while also allowing for a reduced interaction range which substantially reduces calculation time. It is general and can be applied to a wide range of other materials, enabling accurate large-scale simulations of material properties in the presence of different defects.en
dc.description.versionPeer revieweden
dc.format.extent11
dc.format.extent1-11
dc.format.mimetypeapplication/pdfen_US
dc.identifier.citationSchattauer, C, Todorović, M, Ghosh, K, Rinke, P & Libisch, F 2022, ' Machine learning sparse tight-binding parameters for defects ', npj Computational Materials, vol. 8, no. 1, 116, pp. 1-11 . https://doi.org/10.1038/s41524-022-00791-xen
dc.identifier.doi10.1038/s41524-022-00791-xen_US
dc.identifier.issn2057-3960
dc.identifier.otherPURE UUID: e39480d0-dfa1-49be-b689-1315fcf08356en_US
dc.identifier.otherPURE ITEMURL: https://research.aalto.fi/en/publications/e39480d0-dfa1-49be-b689-1315fcf08356en_US
dc.identifier.otherPURE LINK: http://www.scopus.com/inward/record.url?scp=85130340690&partnerID=8YFLogxKen_US
dc.identifier.otherPURE FILEURL: https://research.aalto.fi/files/86360804/Machine_learning_sparse_tight_binding_parameters_for_defects.pdfen_US
dc.identifier.urihttps://aaltodoc.aalto.fi/handle/123456789/115959
dc.identifier.urnURN:NBN:fi:aalto-202208104781
dc.language.isoenen
dc.publisherNature Publishing Group
dc.relation.ispartofseriesnpj Computational Materialsen
dc.relation.ispartofseriesVolume 8, issue 1en
dc.rightsopenAccessen
dc.titleMachine learning sparse tight-binding parameters for defectsen
dc.typeA1 Alkuperäisartikkeli tieteellisessä aikakauslehdessäfi
dc.type.versionpublishedVersion

Files