"Data-driven inverse design of functional materials"
Wes Reinhart, Siemens Corporate Technology (Reinhart.email@example.com)
More than half of the Grand Challenges identified by the National Academy of Engineering involve the design, manufacture, or maintenance of advanced materials, whose properties are inherently derived from their internal structure. This relationship between structure and function is challenging to understand and even harder to predict because it is non-linear, high-dimensional, and results from physical phenomena at many length scales. While conventional materials design relies on human intuition to interpret patterns in known structure-function pairs and infer new materials with similar properties, my work capitalizes on recent advances in both high-performance computer simulation and emerging data-driven methods to enable true “inverse design” of material micro-structure without relying on human expertise.
To do this, I use lessons learned from the traditional “forward design” approach to develop efficient and robust inverse design workflows based on both physics-based and data-driven models, making use of GPU-accelerated computing, hybrid simulation methods accelerated by machine learning, and generative models which require no simulation at all. In my talk, I will discuss how techniques borrowed from data science and machine learning, including the increasingly popular deep learning methods, can be applied to problems in materials science, but I will also highlight the continued role of high-performance physics simulation in predicting the thermodynamic, electromagnetic, and mechanical responses of engineered materials. I will specifically address recent successes I have had applying these hybrid approaches to materials characterization and design problems which should be broadly applicable to the development of improved functionalities and entirely new emergent properties in a wide variety of materials applications.
Wesley Reinhart received his B.S. in Chemical Engineering from the University of Minnesota in 2014. He attended Princeton University on a National Defense Science & Engineering Graduate Fellowship, where he worked on strategies for predicting, under-standing, and controlling colloidal crystallization using large-scale computer simulations and machine learning methods. After earning his PhD in Chemical and Biological Engineering in early 2019, he took his current position of Research Scientist at Siemens Corporate Technology. His current research initiatives there focus on understanding and exploiting the relationship between material structure and function using a combination of data science and high-performance simulation. He has authored 12 journal papers, 5 patents, and currently serves as Principal Investigator for a major DARPA research project.