A graphical representation of a neural network overlayed on a photo of a rack-mounted computers.

An interdisciplinary team of Whiting School faculty members has been chosen to receive a U.S. Department of Energy Award.

This four-year, $4.8 million grant from the Department of Energy’s Office of Science promotes research on scientific machine learning focused on greater predictive capabilities for scientific simulations.

The team is led by Michael Shields and includes Dimitris Giovanis, Somdatta Goswami (soon joining WSE as a faculty member), Lori Graham-BradyYannis Kevrekidis, and Tamer Zaki.

Their project, “Physics and Uncertainty Informed Latent Operator Learning,” seeks to address two of the primary challenges to widespread adoption of scientific machine learning (SciML) methods and their applications in the physical, natural, and engineering sciences: scaling of SciML methods to large-scale problems with highly complex physics and the simultaneous quantification of uncertainty. 

It aims to develop novel physics-informed neural operators that exploit the underlying low-dimensional structure of high-dimensional physics-based models which will focus on applications in fracture mechanics of additively manufactured composites and high-speed fluid flow.