Inference on Galaxy Cluster Fields

During the Summer of 2024, I had the pleasure of working as a Research Assistant at the Department of Statistcs, University of Michigan under the supervision of Professors Jeffrey Regier and Camille Avestruz (Department of Physics). I primarily worked on developing software to generate hyper-realistic astronomical images of galaxy clusters using GalSim with an aim to eventually detect galaxy clusters in real astronomical images. Our work follows the methodology of forward amortized variational inference (FAVI) to learn the parameters that best fit the generative prior distribution. Here is a deeper look into my work.