Machine-Learning Recognition of Dzyaloshinskii-Moriya Interaction from Magnetometry

The Dzyaloshinskii-Moriya interaction (DMI), which is the antisymmetric part of the exchange interaction between neighboring local spins, winds the spin manifold and can stabilize non-trivial spin textures in topological magnets. Since topology is a robust information carrier, characterization techniques that can extract the DMI magnitude are important for the discovery and optimization of spintronic materials. Existing experimental techniques for determining the DMI, such as high-resolution imaging of spin textures and measurement of magnon or transport properties, are time consuming and require access to specialized instrumentation. Here we show that a convolutional neural network (CNN) can extract the DMI magnitude from the minor hysteresis loops, or magnetic "fingerprints," of a material. These hysteresis loops are typically obtained by magnetometry measurements, the most accessible characterization technique for magnets. Our customized CNN was able to accurately estimate the value of the DMI of samples with featureful magnetic fingerprints and provide a confidence level for its estimate. This provides a convenient tool to search for topological spin textures for next-generation information processing.