Learning the Universe

Learning the Universe is a Simons Collaboration with the following goals: 1) develop physically grounded models for galaxy evolution using multiscale simulations 2) build new tools to create detailed ‘synthetic observations’ based on simulations 3) develop machine learning-based techniques to accelerate forward modeling of cosmological galaxy formation simulations 4) incorporate these new accelerated forward models into traditional and implicit likelihood inference frameworks and use observational data to constrain cosmological parameters, the initial conditions of the Universe, and the uncertainties of galaxy formation modeling.