Physics-based machine learning with intrusive scalable Gaussian processes at electrochemical interfaces and beyond (ZOOM ONLY)

Abstract: Electrochemical interfaces and surfaces in the solid state frequently control the behavior of critical reactions and conduction taking place at those surfaces. The calculation of surface behavior given certain fundamental characteristics from the bulk properties of ion conducting materials has always been of paramount interest for this reason. Experimental observations showing that dilute-case theories are not relevant to materials even mildly concentrated in mobile charge carriers continue to pile up, but researchers continue to use them, probably for lack of a utilitarian alternative. Could data science yield a solution? We will discuss a unique, scalable Gaussian process that might hold the key.

Bio: David Mebane received a Ph.D. in Materials Science and Engineering from Georgia Tech in 2007, where he constructed data-driven models of mixed conducting solid oxide fuel cell electrodes. He held postdoctoral fellowships at the Max Planck Institute for Solid State Research in Stuttgart and the National Energy Technology Laboratory in Morgantown before joining the faculty at WVU, where he is currently an Associate Professor in Mechanical and Aerospace Engineering. His work is funded by the National Science Foundation, the U.S. Department of Energy, the American Chemical Society and companies including Process Systems Enterprise and Autodesk. His work entails arranging the marriage of first principles theory and data in models of devices and systems that will enable the global energy transition.

Zoom Link: https://stanford.zoom.us/j/92153920201?pwd=YW5PV1kxek9Cd2xuY0xwWU9zNWdWUT09

Zoom Password: 257509