pop-cosmos: Redshifts and physical properties of KiDS-1000 galaxies
The episode discusses the paper "pop-cosmos: Redshifts and physical properties of KiDS-1000 galaxies." Hosts explore how this study applies Bayesian inference to millions of sources to determine galaxy physical properties, moving beyond simple counting. They detail findings on dusty star-forming contaminants and discuss the technical improvements using GPU acceleration and generative models that allow for scalable, physically motivated analysis.