Boxuan Zhao

Hypothesis-and-Refinement Learning of Organic Structures from Multimodal Spectroscopic Data

Determining molecular structures from spectroscopic data remains fundamentally challenging because the inverse problem is intrinsically underdetermined: individual spectra are sparse, low-dimensional, and encode only partial structural evidence relative to the vast space of possible molecules. We address this challenge...