One of the most pressing issues in theoretical cosmology today is how to optimally interpret the data from future observations of the large-scale structure of the Universe. In this talk, I will discuss a framework based on cosmological N-body simulations that provides accurate predictions as a function of cosmological parameters and astrophysics. The method is a result of several recent advances in computational cosmology regarding the modelling of dark matter, gas, and baryonic effects, galaxy formation, and the development ofneural networks that enable the delivery of theory predictions in a fraction of a second. The approach I will discuss has the potential to significantly increase the amount of cosmological information extracted from surveys like EUCLID, and thus about the nature of dark energy, dark matter, and the mass of neutrinos.
