Conditional Normalising Flows

Using a super-smart puzzle solver to figure out which light maps are likely to come from specific galaxies with certain redshifts.

Technical definition

A deep learning approach that estimates probability distributions over the latent space by normalizing the data through sequential transformations.