SPECIAL SEMINAR: Materials Science in the Information Age

22 Mar
Materials Science Seminars
Dogus Cubuk
Wednesday, March 22, 2017 -
11:00am to 12:00pm
Pierce Hall 209

The study and design of new materials have always benefited from previous measurements and calculations available in literature. With the recent availability of large databases and ever-increasing computational resources, incorporating all relevant data is becoming impossible with traditional tools. I will talk about my recent work on utilizing machine learning to study complex materials phenomena and design new energy materials. I will mainly focus on data-driven, unified models for kinetic and mechanical properties of disordered solids, which span over 7 orders of magnitude in constituent size and have vastly different inter-particle interactions, from atoms to grains.