MetalFinder is a tree-based supervised learning algorithm which can predict the mono-metallic nanoparticle (MMNP) from a Pair Distribution Function (PDF).
Currently MetalFinder is limited to MMNPs with up to 200 atoms of the 7 different structure types: Cubic (sc), body-centered cubic (bcc), face-centered cubic (fcc), hexagonal closed packed (hcp), decahedral, icosahedral and octahedral
Follow these step if you want to train MetalFinder and predict with MetalFinder locally on your own computer.
See the install folder.
To simulate the data used to train MetalFinder open the file:
jupyter notebook 1_Simulate_Data.ipynb
Follow the instructions in the 1_Simulate_Data.ipynb file.
To download the xyz-files (atomic models) and PDF dataset instead of simulating both locally, see the data folder.
To train your own MetalFinder model simply run:
jupyter notebook 2_Training.ipynb
Follow the instructions in the 2_Training.ipynb file. To download the model used for prediction in the article, instead of training at your own computer, see the data folder.
To predict a MMNP using MetalFinder or your own model on a PDF:
jupyter notebook 3_Testing.ipynb
Follow the instructions in the 3_Testing.ipynb file.
Andy S. Anker1
Emil T. S. Kjær1
Marcus N. Weng1
Simon J. L. Billinge2, 3
Raghavendra Selvan4, 5
Kirsten M. Ø. Jensen1
1 Department of Chemistry and Nano-Science Center, University of Copenhagen, 2100 Copenhagen Ø, Denmark.
2 Department of Applied Physics and Applied Mathematics Science, Columbia University, New York, NY 10027, USA.
3 Condensed Matter Physics and Materials Science Department, Brookhaven National Laboratory, Upton, NY 11973, USA.
4 Department of Computer Science, University of Copenhagen, 2100 Copenhagen Ø, Denmark.
5 Department of Neuroscience, University of Copenhagen, 2200, Copenhagen N.
Should there be any question, desired improvement or bugs please contact us on GitHub or through email: ansoan@dtu.dk or etsk@chem.ku.dk.
If you use our code or our results, please consider citing our paper. Thanks in advance!
@article{kjær2022DeepStruc,
title={DeepStruc: Towards structure solution from pair distribution function data using deep generative models},
author={Emil T. S. Kjær, Andy S. Anker, Marcus N. Weng, Simon J. L. Billinge, Raghavendra Selvan, Kirsten M. Ø. Jensen},
year={2022}}
This project is licensed under the Apache License Version 2.0, January 2004 - see the LICENSE file for details.