Presentation Information
[9a-E310-3]Machine learning-based estimation of magnetic parameters from inhomogeneous magnetic domain images
〇Shu Hashimoto1, Yoshinobu Nakatani2, Hiroyuki Awano1, Kenji Tanabe1 (1.Toyota Tech. Inst., 2.UEC)
Keywords:
parameter estimation,magnetic domain,machine learning
Some parameters of magnetic thin films are difficult or time-consuming to measure. For this reason, research is underway on methods that use machine learning to estimate these parameters from images of maze-like magnetic domains. We have previously demonstrated that it is possible to estimate the variance σ of the magnetic anisotropy constant with high accuracy in heterogeneous systems where σ is present. In this study, we focus on homogeneous systems where σ is deliberately excluded, and we investigate and report on the estimation limits and regions where estimation is difficult due to the absence of image features.