Title: From neural nets to cosmic webs: Reconstructing the large scale structure of the universe from discrete tracers.

Abstract: Galaxy formation and evolution cannot be studied independently of large-scale structure. Statistical measurements in cosmological simulations have shown that galaxy evolution depends heavily on the environments they evolve in; moreover, observations at low redshift (z < 0.3) provide evidence that the anisotropic large scale environment impacts at least some galaxy properties. Nevertheless, observing and labeling these environments at higher redshift remains a challenge, leaving us with an incomplete cosmological context of the galaxies we observe. The goal of my PhD project consists of developing a deep learning model capable of reconstructing the gas density field in cosmological volumes using discrete observable tracers such as galaxies. During my talk I will be motivating the applications of this model, presenting the methodology behind it and my current results.