Toward a New Census of the 0 < z < 20  Universe

Current and upcoming surveys conducted by the James Webb Space Telescope and the Rubin Observatory promise to transform the systematic study of galaxy evolution. In this talk I will present our recent works to address the new challenges created by these new data. With the launch of JWST, for the first time, it becomes necessary to allow redshift solutions for the full range of 0 < z < 20. I will demonstrate the utility of informative priors encoding previous observations of galaxies across cosmic time to attain accurate joint posteriors of redshift and key stellar populations metrics. Some successful applications on early JWST observations of this new model, "Prospector-beta", will also be discussed. Meanwhile, the Rubin Observatory, expecting to observe 5 billion galaxies, will map the pathways of galaxy assembly to an incredibly high resolution. However, at >10h/fit, the traditional methods to interpret galaxy imaging on the full data-set require ~100 billion CPU-hours. This is prohibitively expensive. I will present "SBI++" -- a flexible, fast likelihood-free inference tool customized for astronomical applications. It has an extremely fast inference speed of ~1 sec for objects in the training set distribution, and additionally permits parameter inference outside of the trained noise and data at ~1 min per object. SBI++ opens the door for large-scale applications of high-dimensional galaxy population models. This will establish new, in-depth censuses of the cosmic history.