Dr. Soumak Maitra

A central goal of cosmology is to understand the evolution of structures in the Universe across cosmic time. I frame this endeavor as "Charting the Cosmic History": building on global averages and two-point statistics, while extending to higher-order correlations of cosmic tracers that yield more informative descriptions of matter and ionization fields. At low and intermediate redshifts, Lyman-a forest sightlines through the intergalactic medium and its three-point correlation offer an indication of non-Gaussianity in the density field beyond the reach of power spectrum. As we move to higher redshifts and approach the epoch of reionization, the focus has traditionally been on constraining the mean neutral hydrogen fraction. However, the spatial distribution of Lyman-a emitters, whose visibility is regulated by the surrounding ionized regions, naturally encodes information on the sizes and morphology of ionized bubbles, and hence on the topology of reionization. In this talk, I will present results on characterizing cosmic structures beyond two-point correlations across these regimes, and demonstrate how modern machine-learning methods enable more informative reconstructions of density and ionization fields by incorporating complex physics that is difficult to capture with traditional statistical approaches.