From Event Detection to Physical Hypothesis Learning via Automated and Autonomous Microscopy

Electron and scanning probe microscopies are by now the mainstay of areas ranging from fundamental condensed matter physics and materials science to chemistry and biology. In many cases, these imaging methods provide quantitative or semiquantitative information on materials structure and functionalities and allow for the exploration of broad range of responses to global (temperature, chemistry) and local (beam exposure, probe bias and pressure) stimuli. However, the connection between microscopy imaging, spectral data, and specific mechanisms of physical and chemical phenomena often requires extensive studies including multiple human-driven stages.