Determination of the Dzyaloshinskii-Moriya interaction using pattern recognition and machine learning

Abstract

Machine learning is applied to a large number of modern devices that are essential in building an energy-efficient smart society. Audio and face recognition are among the most well-known technologies that make use of such artificial intelligence. In materials research, machine learning is adapted to predict materials with certain functionalities, an approach often referred to as materials informatics. Here, we show that machine learning can be used to extract material parameters from a single image obtained in experiments. The Dzyaloshinskii–Moriya (DM) interaction and the magnetic anisotropy distribution of thin-film heterostructures, parameters that are critical in developing next-generation storage class magnetic memory technologies, are estimated from a magnetic domain image. Micromagnetic simulation is used to generate thousands of random images for training and model validation. A convolutional neural network system is employed as the learning tool. The DM exchange constant of typical Co-based thin-film heterostructures is studied using the trained system: the estimated values are in good agreement with experiments. Moreover, we show that the system can independently determine the magnetic anisotropy distribution, demonstrating the potential of pattern recognition. This approach can considerably simplify experimental processes and broaden the scope of materials research.

Large-Scale Statistical Learning for Mass Transport Prediction in Porous Materials Using 90,000 Artificially Generated Microstructures

“…Note that conventional ANNs learn to perform nonlinear regression using predefined descriptors, whereas CNNs perform their own descriptor extraction directly from the microstructure, expressed as nonlinear compositions of convolution filters. These are then used as input to a conventional ANN that performs the regression (Kawaguchi et al, 2021). CNNs have been used for predicting both permeability (Srisutthiyakorn, 2016;Wu et al, 2018;Araya-Polo et al, 2019;Sudakov et al, 2019;Graczyk and Matyka, 2020;Kamrava et al, 2020) and effective diffusivity (Wu et al, 2019;Wang et al, 2020), although in many cases with small datasets and/or only with 2D structures.…”

Topological characterization of dynamic chiral magnetic textures using machine learning

“…The custom architecture used here is similar to that of LeNet-5 [23]. A comparable custom architecture has been used for other analyses of magnetic systems [12,15]. The network structure was chosen after varying the number of convolutional and fully connected layers, the filter and kernel size, the number of neurons, and the activation function.…”

“…Particularly, machine learning algorithms have been used for finding [10], recognizing [11], and classifying [12] ground states of systems with Dzyaloshinskii-Moriya interactions (DMI) and for reconstructing the spin configuration from data obtained in reciprocal space [13]. Machine learning has also been used to define phase transitions in Ising-like spin systems [14] and to estimate the DMI parameters from images of magnetic domains [15,16]. In contrast to these investigations of statical properties of skyrmionic systems, a recent study [17] applied neural networks for deep learning of skyrmionic dynamical phases from videos.…”

Understanding the Magnetic Microstructure through Experiments and Machine Learning Algorithms

“…In this paper, we are going to address the application of the convolutional neural network (CNN) to understand the modifications of magnetic domains in a perpendicularly magnetized multilayer, which has been observed experimentally by using ion-beam irradiation. Of late, advanced machine learning techniques have acquired immense importance in interdisciplinary research, such as in microstructure optimization, prediction of a magnetic field, phase transition, magnetic grain size study, modeling magnetic domains, , relation between different magnetic chiral states, prediction of effective magnetic spin configurations, , 2D metal–organic frameworks with high magnetic anisotropy, and different components of Hamiltonian including the Dzyaloshinskii–Moriya interaction (DMI), using different deep learning and machine learning methods. From the point of view of atomistic magnetism, researchers , have tried to estimate and analyze various components of Hamiltonian, such as exchange constant, anisotropy constant, and DMI, using different CNNs .…”