Applying Machine Learning to Elucidate Ultrafast Demagnetization Dynamics in Ni and Ni80Fe20

Abstract

Understanding the correlation between fast and ultrafast demagnetization processes is crucial for elucidating the microscopic mechanisms underlying ultrafast demagnetization, which is pivotal for various applications in spintronics. Initial theoretical models attempted to establish this correlation but faced challenges due to the complex interplay of physical phenomena. To address this, we employed a variety of machine learning methods, including supervised learning regression algorithms and symbolic regression, to analyze limited experimental data and derive meaningful mathematical expressions between demagnetization time and the Gilbert damping factor. The results reveal that polynomial regression and K-nearest neighbors algorithms perform best in predicting demagnetization time. Additionally, sure-independence-screening-and-sparsifying-operator (SISSO) as a symbolic regression method suggested a direct correlation between demagnetization time and damping factor for Ni and Ni80Fe20, indicating spin-flip scattering predominantly influences the ultrafast demagnetization mechanism. The developed models demonstrate promising predictive capabilities, validated against independent experimental data. Comparative analysis between different materials underscores the significant impact of material properties on ultrafast demagnetization behavior. This study underscores the potential of machine learning in unraveling complex physical phenomena and offers valuable insights for future research in ultrafast magnetism.