As Python-based APIs from major equipment manufacturers are becoming increasingly available, automated experimentation is now accessible to a broad range of scientists in both academia and industry. This webinar is tailored to meet these evolving needs. It is specifically designed to bridge the gap between theoretical data science concepts and their applications in automated microscopy, focusing on developing practical experimental workflows. The course provides an in-depth physicist's perspective on Bayesian methods and includes introduction into the GPax library. This approach is aimed at developing both intuition and practical skills necessary for applying Bayesian optimization in real-world experimental settings. Building upon optimization, we will discuss the construction of multistep workflows targeted at discovery of structure-property relationships, identifying physical laws, co-navigation between theory and experimental spaces, and co-orchestration of several tools. By the end of the course, attendees are expected to have a comprehensive understanding of how to apply Bayesian optimization methods effectively for simple tasks such as image optimization, and extend them to real world physics and materials science problems, making this course an invaluable resource for those looking to harness the full potential of ML/AI for modern experimental science.

Webinar Objectives:

• Introduce the general concept of reward-driven workflow design in automated microscopy.
• Provide the basic introduction into Gaussian Processes (GP) and GP-based Bayesian Optimization (BO)
• Introduce structured Gaussian Processes (sGP) and sGP-BO as a framework combining physics discovery and data science.
• Present deep kernel learning (DKL) as an active learning method for structure-property discovery
• Discuss human in the loop automated experiment workflows (hAE) and accelerated microscopy
• Introduce extensions of multi-task and multi-objective GP to design co-orchestration of multiple tools.
• Discuss non-myopic workflows in data analysis and automated experiment.

Sergei V. Kalinin is a Weston Fulton Chair Professor at the University of Tennessee, Knoxville. In 2022 – 2023, he has been a Principal Scientist at Amazon (special projects). Before then, he had spent 20 years at Oak Ridge National Laboratory where he was Corporate Fellow and Group Leader at the Center for Nanophase Materials Sciences. He received his MS degree from Moscow State University in 1998 and Ph.D. from the University of Pennsylvania (with Dawn Bonnell) in 2002. He has >15 years of experience applying AI in experimental physical sciences from data analytics, theory-experiment matching, and automated and autonomous microscopy and materials synthesis for applications including physics discovery, atomic fabrication, as well as mesoscopic studies of electrochemical, ferroelectric, and transport phenomena via scanning probe microscopy. Sergei has co-authored >650 publications, with a total citation of ~50,000 and an h-index of >112. He is a fellow of AAIA, MRS, APS, IoP, IEEE, Foresight Institute, and AVS; a recipient of the Medard Welch Award of AVS (2023), Feynmann Prize of Foresight Institute (2022), Blavatnik Award for Physical Sciences (2018), RMS Medal for Scanning Probe Microscopy (2015), Presidential Early Career Award for Scientists and Engineers (PECASE) (2009); Burton Medal of Microscopy Society of America (2010); 5 R&D100 Awards (2008, 2010, 2016, 2018, and 2023); and a number of other distinctions.

This webinar is intended for researchers, students, technologists and others involved or interested in using machine learning to enable practically relevant automated experiment workflows in electron and scanning probe microscopy. The webinar will be valuable for a large audience, from young scientists to engineers familiar with microscopies, and assumes no prior knowledge of ML/AI methods.  

Each webinar includes a link to a pdf file of the copyrighted webinar notes. During the webinar you may ask questions. Presenter(s) will review and answer them periodically throughout the presentation. 

You must be present the day of the webinar, recorded versions will not be available.