Deep Learning and Computational Physics

Introduction.- Introduction to deep neural networks.- Residual neural networks.- Convolutional Neural Networks.- Solving PDEs with Neural Networks.- Operator Networks.- Generative Deep Learning.

Assad Oberai is the Hughes Professor of Aerospace and Mechanical Engineering in the Viterbi School of Engineering. He earned a Bachelor of Engineering degree from Osmania University, an MS from the University of Colorado, and a PhD from Stanford University all in Mechanical Engineering.  He has held academic appointments at Boston University, Rensselaer Polytechnic Institute, and the University of Southern California.  Assad leads a group that designs, implements, and applies data- and physics-based models and algorithms to solve problems in engineering and science. Problems such as better detection, diagnosis, and care of diseases like cancer, understanding the role of mechanics and physics in medicine and biology, modeling the evolution of multi-physics and multiscale systems, and reduced-order models for aerospace and mechanical systems. Assad is a Fellow of the American Academy of Mechanics, American Society of Mechanical Engineers, the American Institute of Medical and Biological Engineering, and the United States Association of Computational Mechanics. 

Deep Ray is an Assistant Professor of Mathematics at the University of Maryland. He earned his Bachelor of Mathematics from University of Delhi, followed by a Masters and PhD in Mathematics from Tata Institute of Fundamental Research - Center for Applicable Mathematics. He has held research positions at ETH Zurich, EPFL, Rice University and University of Southern California. Deep's research lies at the interface of conventional numerical analysis and machine learning. He focuses on identifying computational bottlenecks in existing numerical algorithms and resolving them by the careful integration of machine learning tools. He has used such techniques to design efficient shock-capturing methods, build deep learning-based surrogate models to solve partial differential equations, develop differentiable models for constrained optimization, and solve Bayesian