Extracting structural motifs from pair distribution function data of nanostructures using explainable machine learning

作     者：Andy S.Anker Emil T.S.Kjær Mikkel Juelsholt Troels Lindahl Christiansen Susanne Linn Skjærvø Mads Ry Vogel Jørgensen Innokenty Kantor Daniel Risskov Sørensen Simon J.L.Billinge Raghavendra Selvan Kirsten M.Ø.Jensen 

作者机构：Department of Chemistry and Nano-Science CenterUniversity of Copenhagen2100CopenhagenDenmark Department of MaterialsUniversity of OxfordParks RoadOxfordUK Department of Chemistry&iNANOAarhus University8000AarhusDenmark MAX IV LaboratoryLund University22484LundSweden Department of PhysicsTechnical University of Denmark2880LyngbyDenmark Department of Applied Physics and Applied MathematicsColumbia UniversityNew YorkNY10027USA Condensed Matter Physics and Materials Science DepartmentBrookhaven National LaboratoryUptonNY11973USA Department of Computer ScienceUniversity of Copenhagen2100CopenhagenDenmark Department of NeuroscienceUniversity of Copenhagen2200CopenhagenDenmark 

出 版 物：《npj Computational Materials》 (计算材料学（英文）)

年 卷 期：2022年第8卷第1期

页      面：2053-2063页

核心收录：

学科分类：07[理学] 070202[理学-粒子物理与原子核物理] 0702[理学-物理学] 

基　　金：This work is part of a project that has received funding from the European Research Council(ERC)under the European Union’s Horizon 2020 Research and Innovation Programme(grant agreement No.804066) We are grateful to the Villum Foundation for financial support through a Villum Young Investigator grant(VKR00015416) Funding from the Danish Ministry of Higher Education and Science through the SMART Lighthouse is gratefully acknowledged.We furthermore thank DANSCATT(supported by the Danish Agency for Science and Higher Education)for support This research used resources of the Advanced Photon Source,a US Department of Energy(DOE)Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under contract No.DE-AC02-06CH11357 We acknowledge MAX IV Laboratory for time on Beamline DanMAX under Proposal 20200731 Research conducted at MAX IV is supported by the Swedish Research council under contract 2018-07152 the Swedish Governmental Agency for Innovation Systems under contract 2018-04969 Formas under contract 2019-02496.DanMAX is funded by the NUFI grant no.4059-00009 B.S.J.L.B.was supported by the U.S.National Science Foundation through grant DMREF-1922234 

主　　题：explain structural meaningful 

摘      要：Characterization of material structure with X-ray or neutron scattering using *** Distribution Function(PDF)analysis most often rely on refining a structure model against an experimental ***,identifying a suitable model is often a ***,automated approaches have made it possible to test thousands of models for each dataset,but these methods are computationally expensive and analysing the output,*** structural information from the resulting fits in a meaningful way,is *** Machine Learning based Motif Extractor(ML-MotEx)trains an ML algorithm on thousands of fits,and uses SHAP(SHapley Additive exPlanation)values to identify which model features are important for the fit *** use the method for 4 different chemical systems,including disordered nanomaterials and ***-MotEx opens for a type of modelling where each feature in a model is assigned an importance value for the fit quality based on explainable ML.