Efficient few-shot machine learning for classification of EBSD patterns

Abstract Deep learning is quickly becoming a standard approach to solving a range of materials science objectives, particularly in the field of computer vision. However, labeled datasets large enough to train neural networks from scratch can be challenging to collect. One approach to accelerating th...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Kevin Kaufmann, Hobson Lane, Xiao Liu, Kenneth S. Vecchio
Formato: article
Lenguaje:EN
Publicado: Nature Portfolio 2021
Materias:
R
Q
Acceso en línea:https://doaj.org/article/b1148711719344a495d8dd02d6386f09
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:b1148711719344a495d8dd02d6386f09
record_format dspace
spelling oai:doaj.org-article:b1148711719344a495d8dd02d6386f092021-12-02T14:26:12ZEfficient few-shot machine learning for classification of EBSD patterns10.1038/s41598-021-87557-52045-2322https://doaj.org/article/b1148711719344a495d8dd02d6386f092021-04-01T00:00:00Zhttps://doi.org/10.1038/s41598-021-87557-5https://doaj.org/toc/2045-2322Abstract Deep learning is quickly becoming a standard approach to solving a range of materials science objectives, particularly in the field of computer vision. However, labeled datasets large enough to train neural networks from scratch can be challenging to collect. One approach to accelerating the training of deep learning models such as convolutional neural networks is the transfer of weights from models trained on unrelated image classification problems, commonly referred to as transfer learning. The powerful feature extractors learned previously can potentially be fine-tuned for a new classification problem without hindering performance. Transfer learning can also improve the results of training a model using a small amount of data, known as few-shot learning. Herein, we test the effectiveness of a few-shot transfer learning approach for the classification of electron backscatter diffraction (EBSD) pattern images to six space groups within the $$\left( {4/m \overline {3} 2/m} \right)$$ 4 / m 3 ¯ 2 / m point group. Training history and performance metrics are compared with a model of the same architecture trained from scratch. In an effort to make this approach more explainable, visualization of filters, activation maps, and Shapley values are utilized to provide insight into the model’s operations. The applicability to real-world phase identification and differentiation is demonstrated using dual phase materials that are challenging to analyze with traditional methods.Kevin KaufmannHobson LaneXiao LiuKenneth S. VecchioNature PortfolioarticleMedicineRScienceQENScientific Reports, Vol 11, Iss 1, Pp 1-12 (2021)
institution DOAJ
collection DOAJ
language EN
topic Medicine
R
Science
Q
spellingShingle Medicine
R
Science
Q
Kevin Kaufmann
Hobson Lane
Xiao Liu
Kenneth S. Vecchio
Efficient few-shot machine learning for classification of EBSD patterns
description Abstract Deep learning is quickly becoming a standard approach to solving a range of materials science objectives, particularly in the field of computer vision. However, labeled datasets large enough to train neural networks from scratch can be challenging to collect. One approach to accelerating the training of deep learning models such as convolutional neural networks is the transfer of weights from models trained on unrelated image classification problems, commonly referred to as transfer learning. The powerful feature extractors learned previously can potentially be fine-tuned for a new classification problem without hindering performance. Transfer learning can also improve the results of training a model using a small amount of data, known as few-shot learning. Herein, we test the effectiveness of a few-shot transfer learning approach for the classification of electron backscatter diffraction (EBSD) pattern images to six space groups within the $$\left( {4/m \overline {3} 2/m} \right)$$ 4 / m 3 ¯ 2 / m point group. Training history and performance metrics are compared with a model of the same architecture trained from scratch. In an effort to make this approach more explainable, visualization of filters, activation maps, and Shapley values are utilized to provide insight into the model’s operations. The applicability to real-world phase identification and differentiation is demonstrated using dual phase materials that are challenging to analyze with traditional methods.
format article
author Kevin Kaufmann
Hobson Lane
Xiao Liu
Kenneth S. Vecchio
author_facet Kevin Kaufmann
Hobson Lane
Xiao Liu
Kenneth S. Vecchio
author_sort Kevin Kaufmann
title Efficient few-shot machine learning for classification of EBSD patterns
title_short Efficient few-shot machine learning for classification of EBSD patterns
title_full Efficient few-shot machine learning for classification of EBSD patterns
title_fullStr Efficient few-shot machine learning for classification of EBSD patterns
title_full_unstemmed Efficient few-shot machine learning for classification of EBSD patterns
title_sort efficient few-shot machine learning for classification of ebsd patterns
publisher Nature Portfolio
publishDate 2021
url https://doaj.org/article/b1148711719344a495d8dd02d6386f09
work_keys_str_mv AT kevinkaufmann efficientfewshotmachinelearningforclassificationofebsdpatterns
AT hobsonlane efficientfewshotmachinelearningforclassificationofebsdpatterns
AT xiaoliu efficientfewshotmachinelearningforclassificationofebsdpatterns
AT kennethsvecchio efficientfewshotmachinelearningforclassificationofebsdpatterns
_version_ 1718391389891330048