Dr. Ziatdinov’s research is directed primarily toward the synergy of machine learning, experiment, and theory to accelerate discoveries in physical sciences. This includes the development of science-informed machine learning workflows capable of incorporating prior domain knowledge, the establishment of critical links between cutting-edge instrumental platforms and high-performance computing facilities, and the enablement of the on-the-fly analysis of streaming data for feedback and instrument control. One of his current main interests is to transform electron and scanning probe microscopy platforms at ORNL into autonomous systems for scientific discovery. Dr. Ziatdinov is a creator of several widely used in the experimental community open-source software packages, including AtomAI for deep and machine learning applications in microscopy, pyroVED for applications of invariant autoencoders in the image and spectral analysis, and GPax for physics-based active learning and Bayesian optimization in automated experiments.
IEEE Spectrum: Self-Driving Microscopes to Navigate the Nanoscale