Mrigank Dhingra

Iterative Refinement Diffusion for Super-Resolved Data Assimilation of Multiscale Physical Systems

Recovering high-resolution states from sparse, low-resolution observations is a central challenge in scientific machine learning and data assimilation. Classical data assimilation exploits temporal information through forecast-analysis cycles, but often requires repeated access to expensive high-resolution forecast mod...