Towards Space Group Determination from EBSD Patterns: The Role of Deep Learning and High-throughput Dynamical Simulations
This research develops a deep learning framework to automatically classify crystal space groups from Electron Backscatter Diffraction (EBSD) patterns using high-throughput dynamical simulations. By training models like Convolutional Neural Networks and Maximum Classifier Discrepancy on simulated data, the method achieves over 90% accuracy, enabling fast determination of crystal structures for newly synthesized nanomaterials.