Improvement of CNN-Based Anisotropic ...
Document type :
Article dans une revue scientifique: Article original
Permalink :
Title :
Improvement of CNN-Based Anisotropic Magnetostatic Field Computation via Adaptive Data Subset Selection
Author(s) :
Gong, Ruohan [Auteur]
Laboratoire d'Électrotechnique et d'Électronique de Puissance (L2EP) - ULR 2697
Tang, Zuqi [Auteur correspondant]
Laboratoire d'Électrotechnique et d'Électronique de Puissance (L2EP) - ULR 2697
Laboratoire d'Électrotechnique et d'Électronique de Puissance (L2EP) - ULR 2697
Tang, Zuqi [Auteur correspondant]
![refId](/themes/Mirage2//images/idref.png)
Laboratoire d'Électrotechnique et d'Électronique de Puissance (L2EP) - ULR 2697
Journal title :
IEEE Transactions on Magnetics
Abbreviated title :
IEEE Trans. Magn.
Volume number :
58
Pages :
1-4
Publisher :
Institute of Electrical and Electronics Engineers (IEEE)
Publication date :
2022-09
ISSN :
0018-9464
English keyword(s) :
Anisotropic magnetic material
convolutional neural network (CNN)
deep learning (DL)
subset selection
convolutional neural network (CNN)
deep learning (DL)
subset selection
HAL domain(s) :
Sciences de l'ingénieur [physics]
English abstract : [en]
A numerical issue arises when we extend the convolutional neural network (CNN) U-net to the anisotropic magnetostatic field computation. The output magnetic field has a significant gradient with respect to the input geometry ...
Show more >A numerical issue arises when we extend the convolutional neural network (CNN) U-net to the anisotropic magnetostatic field computation. The output magnetic field has a significant gradient with respect to the input geometry parameter, which introduces inevitable errors in the training process to degrade the performance of deep learning (DL). To address this issue, the subset selection approach is utilized to divide the whole database into several subsets, where the samples are assigned according to the gradient between the input and output. Then these subsets with different sample densities are combined into a global one. Taking the uniform dataset with the same sample size as a comparison, the influence of subset selection on DL is investigated by comparing the performance of CNN on different datasets. Numerical experiments illustrate that the adaptive subset selection can be employed to improve the accuracy and efficiency of the CNN network.Show less >
Show more >A numerical issue arises when we extend the convolutional neural network (CNN) U-net to the anisotropic magnetostatic field computation. The output magnetic field has a significant gradient with respect to the input geometry parameter, which introduces inevitable errors in the training process to degrade the performance of deep learning (DL). To address this issue, the subset selection approach is utilized to divide the whole database into several subsets, where the samples are assigned according to the gradient between the input and output. Then these subsets with different sample densities are combined into a global one. Taking the uniform dataset with the same sample size as a comparison, the influence of subset selection on DL is investigated by comparing the performance of CNN on different datasets. Numerical experiments illustrate that the adaptive subset selection can be employed to improve the accuracy and efficiency of the CNN network.Show less >
Language :
Anglais
Audience :
Internationale
Popular science :
Non
Administrative institution(s) :
Université de Lille
Centrale Lille
Arts et Métiers Sciences et Technologies
Junia HEI
Centrale Lille
Arts et Métiers Sciences et Technologies
Junia HEI
Research team(s) :
Équipe Outils et Méthodes Numériques
Submission date :
2023-01-25T00:26:17Z
2023-01-25T00:55:33Z
2023-01-27T09:19:51Z
2023-01-25T00:55:33Z
2023-01-27T09:19:51Z