Probabilistic Methods for Data-Driven Reduced-Order Modeling

Explore probabilistic methods for data-driven reduced-order modeling in this comprehensive talk by Mengwu Guo from the University of Twente. Delve into efficient and credible multi-query, real-time simulations crucial for digital twinning. Discover three key approaches: reduced-order surrogate modeling using Gaussian process regression, deep kernel learning for nonlinear dimensionality reduction, and Bayesian reduced-order operator inference. Learn how these methods guarantee improved efficiency and computational credibility through uncertainty quantification. Gain insights into their applications in various contexts of computational engineering and sciences, and understand how they contribute to achieving reliable and efficient simulations of large-scale engineering assets.

DDPS | 'Probabilistic methods for data-driven reduced-order modeling'