In the next decade, the most precise constraints on the parameters of the ΛCDM model (or extensions of this model) will come from the information embedded in the large-scale structure (LSS) of the universe. The next generation of LSS surveys, such as Euclid, will open a window for further testing of theoretical models. However, diving into the analysis of LSS data to obtain cosmological constraints has clear downsides: expensive computational time, complicated recipes for the theoretical observables, the modelling of those observables at small scales and even the hassle of developing a Bayesian likelihood to perform the statistical analysis. Issues multiply if you want to go beyond the LCDM scenario. In this talk, having Euclid as the working-case scenario, I will explain how we can address all these points when working with LSS data, by highlighting the undergoing theoretical work and the (possible) need of using advanced computational techniques (Machine Learning) to obtain the posterior distributions of cosmological parameters.
Local contact & organization : Lisa GOH WAN KHEE, Vilasini TINNANERI SREEKANTH, Ezequiel CENTOFANI
DAp