PhD scholarship in Error Propagation and Implicit Priors in Computational UQ - DTU Compute

Job description

Would you like to work with applied and computational mathematics in a collaborative and ambitious research team? Are you interested in establishing new insight about the influence and impact of data errors in the computed results? We can offer a 3-year PhD position with scientific challenges and international collaboration. Responsibilities and qualifications: This PhD project is concerned with analysis of the influence of data errors in iterative solvers for inverse problems such as computed tomography and image deblurring. It is part of the research project DUDE: Data Uncertainty and Design/Reconstruction Errors funded by Independent Research Fund Denmark. The work is carried out together with two postdocs in the Section for Scientific Computing. The goal is to establish a solid understanding of how data errors propagate in iterative regularization methods such as CGLS and GMRES. In addition, we aim to interpret these methods in the framework of computational uncertainty quantification. The project continues the work in the project CUQI: Computational Uncertainty Quantification for Inverse Problems. We will utilize and expand our software CUQIpy, and the results will be tested o...