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

• Published in: npj computational materials Vol. 8; no. 1; pp. 1 - 11
• Main Authors: Anker, Andy S., Kjær, Emil T. S., Juelsholt, Mikkel, Christiansen, Troels Lindahl, Skjærvø, Susanne Linn, Jørgensen, Mads Ry Vogel, Kantor, Innokenty, Sørensen, Daniel Risskov, Billinge, Simon J. L., Selvan, Raghavendra, Jensen, Kirsten M. Ø.
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 01-10-2022; Nature Publishing Group; Nature Portfolio
• Subjects: 639/301/357/537; 639/638/298; Algorithms; Automation; Bioinformatics (Computational Biology); Bioinformatik (beräkningsbiologi); Characterization and Evaluation of Materials; Chemistry; Chemistry and Materials Science; Computational Intelligence; Computer and Information Science; Data- och informationsvetenskap (Datateknik); Datasets; Distribution functions; Laboratories; Learning algorithms; Machine learning; Materials Science; Mathematical and Computational Engineering; Mathematical and Computational Physics; Mathematical Modeling and Industrial Mathematics; Methods; Molybdenum; Nanomaterials; Nanoparticles; Nanotechnology; Natural Sciences; Naturvetenskap; Neutron scattering; Physics; Structural analysis; Theoretical
• ISSN: 2057-3960; 2057-3960
• DOI: 10.1038/s41524-022-00896-3

Characterization of material structure with X-ray or neutron scattering using e.g. Pair Distribution Function (PDF) analysis most often rely on refining a structure model against an experimental dataset. However, identifying a suitable model is often a bottleneck. Recently, automated approaches have made it possible to test thousands of models for each dataset, but these methods are computationally expensive and analysing the output, i.e. extracting structural information from the resulting fits in a meaningful way, is challenging. Our
M
achine
L
earning based
Mot
if
Ex
tractor (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 quality. We use the method for 4 different chemical systems, including disordered nanomaterials and clusters. ML-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.