Research Group on Computational Uncertainty Quantification

Our research focuses on the mathematical and computational foundations of uncertainty quantification (UQ) for inverse problems, where incomplete, noisy, or indirect data leads to uncertainty in reconstructed solutions. We develop rigorous mathematical frameworks to characterize and analyze uncertainty, with particular emphasis on Bayesian and probabilistic approaches.

On the computational side, we design efficient numerical algorithms and scalable software that enable practical uncertainty-aware inference in high-dimensional and large-scale settings. Our work bridges theory and computation, ensuring that developed methods are both mathematically well-founded and computationally feasible.

These methods are motivated by and applied to real-world problems, including medical imaging, seismic imaging, industrial imaging, and related applications, where reliable decision-making requires not only accurate reconstructions but also quantified confidence in the results.