Robust Equation Discovery and Sparse Sensing with Guarantees

Explore a statistical mechanics approach for robust sparse equation discovery and optimal sensor placement in this one-hour talk by Krithika Manohar from the University of Washington. Learn how free energies can be used to analyze optimization landscapes, optimize hyperparameters, and quantify uncertainty in the presence of noise and limited data. Discover how this perspective adapts to optimal sensor placement, providing insights into optimization landscapes and critical noise regimes for reconstructing high-dimensional fields using sparse sensors and data-driven priors. Gain practical knowledge on applying these tools for constrained optimization of sensor placement in nuclear digital twins. Delve into the crucial task of extracting governing equations and dynamics from data for prediction, sensing, and control of resource-constrained engineering systems.