Learning to See Sharper: A Physics-Informed Artificial Intelligence Framework for Super-Resolving Galaxy Spectra
This work introduces a deep-learning framework to super-resolve low-resolution galaxy spectra by a factor of about 10, improving resolution from R~100 to R~1000. The model uses a three-stage architecture—coarse reconstruction, redshift inference, and fine refinement—trained on JWST data. It successfully enhances the signal and resolves features like emission line doublets that were previously invisible at lower resolutions.