A forward modeling approach to analyzing galaxy clustering with S im BIG

Abstract

'Pablo Lemos' 'Elena Massara' 'Chirag Modi' 'Azadeh Moradinezhad Dizgah' 'Bruno Régaldo-Saint Blancard' 'Muntazir M. Abidi'] Title: Significance The three-dimensional spatial distribution of galaxies encodes key cosmological information on the nature of dark energy and the contents of the Universe. Current analyses of the statistical clustering of galaxies successfully extract information on large scales that are well described by analytic models. They, however, struggle on smaller, nonlinear, scales. Here, we present SimBIG, an approach to galaxy clustering analyses that can extract information on nonlinear regimes by exploiting high-fidelity simulations and inference based on machine learning. To demonstrate its advantages, we apply SimBIG to 109,636 galaxies of the BOSS survey and analyze a standard summary statistic of the galaxy distribution. Our constraints are consistent with previous works and substantially improve their precision on select cosmological parameters.