The machine learning revolution has seemed to bring order to many problems of big data, by detecting patterns and predicting the response of systems that may appear to behave free of natural laws. Successes in learning the behaviors leading to individuals' choices and crowd dynamics, and the excitement around autonomous mobility are just some examples of machine learning's successes in such spheres. There have been successful applications to the search for new materials that bear relations to these "big data" classification problems. However, when it comes to physical systems governed by conservation laws, the role of machine learning has been more controversial, seen as it is to proceed without regard to the underlying physics. In this talk, I will present our recent work in exploring the role of machine learning methods in discovering, or aiding, the search for physics. This will come by way of two case studies. The first is on using machine learning algorithms to represent high-dimensional free energy surfaces with the goal of identifying precipitate morphologies in alloy systems. The second, of potentially greater reach, is in representing emergent behavior to bridge multiscale problems in materials physics.
About the Speaker
Krishna Garikipati obtained his undergraduate degree from the Indian Institute of Technology, Bombay, and his Masters and PhD from Stanford University. The latter in 1996. Since 2000 he has been a faculty member at University of Michigan, where he is now a Professor of Mechanical Engineering, and Mathematics. In 2016, he was appointed Director of the Michigan Institute for Computational Discovery & Engineering, which is the focus of all research education and outreach in computational science and engineering across eight schools, 30 departments and 130 faculty members at the University of Michigan. He is a past recipient of a PECASE award, and the Alexander von Humboldt Fellowship.
Host: Professor Narayan Aluru