Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

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

Anker, Andy S. ; Kjær, Emil T.S. ; Juelsholt, Mikkel ; Christiansen, Troels Lindahl ; Skjærvø, Susanne Linn ; Jørgensen, Mads Ry Vogel LU orcid ; Kantor, Innokenty LU ; Sørensen, Daniel Risskov LU orcid ; Billinge, Simon J.L. and Selvan, Raghavendra , et al. (2022) In npj Computational Materials 8(1).
Abstract

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 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... (More)

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 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 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.

(Less)
Please use this url to cite or link to this publication:
author
; ; ; ; ; ; ; ; and , et al. (More)
; ; ; ; ; ; ; ; ; and (Less)
organization
publishing date
type
Contribution to journal
publication status
published
subject
in
npj Computational Materials
volume
8
issue
1
article number
213
publisher
Nature Publishing Group
external identifiers
  • scopus:85139230506
ISSN
2057-3960
DOI
10.1038/s41524-022-00896-3
language
English
LU publication?
yes
id
6de49230-cb4b-45d0-8f38-9167098f74c6
date added to LUP
2022-12-19 11:56:32
date last changed
2022-12-19 11:56:32
@article{6de49230-cb4b-45d0-8f38-9167098f74c6,
  abstract     = {{<p>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 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 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.</p>}},
  author       = {{Anker, Andy S. and Kjær, Emil T.S. and Juelsholt, Mikkel and Christiansen, Troels Lindahl and Skjærvø, Susanne Linn and Jørgensen, Mads Ry Vogel and Kantor, Innokenty and Sørensen, Daniel Risskov and Billinge, Simon J.L. and Selvan, Raghavendra and Jensen, Kirsten M.Ø.}},
  issn         = {{2057-3960}},
  language     = {{eng}},
  number       = {{1}},
  publisher    = {{Nature Publishing Group}},
  series       = {{npj Computational Materials}},
  title        = {{Extracting structural motifs from pair distribution function data of nanostructures using explainable machine learning}},
  url          = {{http://dx.doi.org/10.1038/s41524-022-00896-3}},
  doi          = {{10.1038/s41524-022-00896-3}},
  volume       = {{8}},
  year         = {{2022}},
}