Simulation-Based Inference of Galaxies (SimBIG) is a framework for extracting cosmological information from the observed 3D distribution of galaxies using state-of-the-art simulations and neural inference. It combines simulations that accurately model the galaxy distribution and the detailed observational effects of galaxy surveys with neural inference based on deep generative models.
SimBIG enables researchers to robustly extract additional cosmological information in higher-order galaxy clustering down to small non-linear scales, currently inaccessible with standard analyses. With this additional information, researchers can place precise constraints on the growth and expansion histories of the universe.
