Classification of Non-Topological Magnetic Configurations Using Machine Learning

Electron microscopy equipment is extensively used in semiconductor manufacturing and research to Figure out different magnetic configurations in materials. Machine learning is used to carry out efficient studies and analyses in the field of condensed matter physics. This paper proposes a machine learning approach that classifies between different non-topological magnetic configurations. We propose multiple machine learning models that are trained on data that has been generated based on the physical properties of respective magnetic structures. Multi-class classification between non-topological structures such as Ferromagnets, Anti-Ferromagnets & Spin-spiral were done to study how machine learning techniques perform on this task. To accomplish this, convolutional neural network (CNN) and support vector machine (SVM) with principal component analysis (PCA) algorithms have been used. Our experimental results show that both CNN and SVM perform exceptionally well in distinguishing between different non-topological magnetic configurations.