UQ for ML and ML for UQ - Uncertainty Quantification and Machine Learning in Physics-Based Modeling

UQ for ML and ML for UQ - Uncertainty Quantification and Machine Learning in Physics-Based Modeling

Explore the interrelated roles of Uncertainty Quantification (UQ) and Machine Learning (ML) in physics-based computational modeling through this 57-minute seminar by Michael D. Shields, Associate Professor of Civil & Systems Engineering at Johns Hopkins University. Delve into the concepts of "UQ for ML" and "ML for UQ" and their significance in modern physics-based computational modeling paradigms. Examine how these approaches are applied in various fields, from multi-scale materials modeling to high energy-density physics. Gain insights into the importance of addressing uncertainties in parameters, inputs, and model structures when developing physics-based models and implementing scientific machine learning methods.