Deep Learning for Microscopy Image Analysis in Materials Science: Advancing Research and Education Workshop

Through this workshop, the organizers aim to bring together researchers who specialize in Transmission Electron Microscopy (TEM) characterization of materials and researchers who specialize in machine learning. The goal is to create an environment where these researchers can collaborate, learn from one another, and share their knowledge and passion for their respective fields. By bringing these groups together, the organizers hope to foster new ideas and insights that can drive future research in the field of  materials science, with state-of-the-art deep learning techniques for microscopy image analysis. Ion and neutron irradiated material microstructures will be highlighted as a valuable model system for exercising advanced machine learning methods for automated microscopy image analysis.

The workshop is designed to span two days. The first day will focus on research, with several renowned speakers discussing topics such as automated experimentation in electron microscopy, radiation defects in nuclear materials, and high-throughput microscopy and machine learning, among others. The second day will be an education day, where attendees can gain hands-on experience with machine learning techniques, such as U-Net, Mask R-CNN, YOLO, and unsupervised learning techniques such as Auto-encoders and Regular Variational Techniques. The workshop will also cover topics such as the use of Google Colab, Anaconda, data augmentation techniques, labeling data, and the final processing steps after obtaining data from the ML algorithm.