| Abstract Scope |
Bringing machine learning to microscopy is a North Star goal for areas ranging from materials and ceramics science to condensed matter physics – with the dream applications of autonomous discovery of structure-property relationships, exploring physics of nanoscale systems, and building matter on nanometer and atomic scales. In this presentation, I illustrate the examples of fully autonomous microscopy systems for exploring grain boundary transport behavior in polycrystalline materials, discovering the origins of non-linearity, nucleation biases, and work of switching in ferroelectric materials, and discovery of the ferroelectric domain growth laws. The same algorithms can be deployed on the electron microscopes for the discovery of the regions of interest based on their EELS signatures. These workflows are further developed to allow for explainable AI and human in the loop interventions. Overall, the development of advanced ML methods now opens a pathway to complete overhaul of classical operational concepts in microscopy, enhancing its potential for scientific discovery across multiple disciplines. |