Application of Machine Learning to Electron Microscopy Data - IPAM at UCLA

Explore a 42-minute conference talk on the application of machine learning to electron microscopy data presented by Ivan Pedro Lobato Hoyos from the University of Antwerp at IPAM's Mathematical Advances for Multi-Dimensional Microscopy Workshop. Delve into recent advancements in electron microscope data acquisition, allowing for the collection of multidimensional sample data at rates of approximately 1TB/hour. Discover how machine learning techniques can be employed to reverse engineer experimental data and retrieve crucial information about specimens, including three-dimensional atomic positions, compositions, and valence states. Examine the mathematical modeling of undistorted and distorted SEM/STEM/TEM data using accurate and fast electron-specimen interaction models. Learn about neural networks capable of compensating for high levels and combinations of distortions in experimental SEM, STEM, TEM, and CBED datasets. Gain insights into the challenges and solutions in extracting valuable information from electron microscopy data using advanced mathematical and machine learning approaches.