Published in: 2018 Progress in Electromagnetics Research Symposium (PIERS-Toyama)
Magnetic materials and their intrinsic properties have been utilized in different types of devices since the past many decades. From a computational perspective, these properties are characterized and various field interactions are visualized with the help of simulation based softwares. These softwares provide a solution for different types of models using well known numerical techniques. Over the past few years, the field of Machine Learning has started to be utilized to supplement numerical based simulations. As a result, simulations are able not only to numerically compute, but also approximate, classify, and forecast various properties and field states. In this article, we discuss the utilization of deep learning in context of identification of various magnetic states and approximation of magnetic fields. Accuracy is reported for various test cases against different types of deep learning architectures, showing good results for some architectures, whereas unacceptable results for others
-
Carrasquilla, J. and R. G. Melko, \Machine learning phases of matter", Nature Physics, Vol. 13, 431{434, 2017.
-
Adly, A. A. and S. K. Abd-el-Ha z, \Utilizing neural networks in magnetic media modeling and eld computation: A review", J. Adv. Res., Vol. 5, No. 6, 615{627, 2014.