Institute for Advanced Study/Princeton University Early Universe/Cosmology Lunch Discussion

Abstract 1: The bispectrum is the leading non-Gaussian statistic in large-scale structure, providing complementary information to the power spectrum. I present the two-loop perturbative description of the bispectrum within the EFTofLSS framework. The UV sensitivity of the two-loop correction in the double-hard region can be absorbed by the four independent EFT operators and parameters known from the one-loop bispectrum. A simplified treatment is used for the single-hard region, with one extra EFT parameter at two-loop order. Comparing the results to N-body data, we find that when going from one- to two-loop, the range of wavenumbers with percent-level agreement extends from 0.08 h/Mpc to 0.15 h/Mpc. Finally, I discuss the impact of using exact kernels instead of Einstein-de-Sitter ones for the power spectrum and one-loop bispectrum.
Abstract 2: I will discuss the recent progress in modeling 1-point probability distribution function (PDF) for the spherically-averaged matter density field. First, I will briefly recap the saddle-point (‘semiclassical’) approximation. Then, I will focus on the computation of aspherical prefactor coming from integration over small fluctuations around the saddle-point solution. This part receives large contributions from unphysical UV modes and must be renormalized just like the 1-loop correction to the power spectrum is renormalized in the EFTofLSS. I will present a new model for renormalization of the aspherical prefactor stemming from the effective long-scale stress-energy tensor written for aspherical short-wavelength perturbations. Finally, I will confront the predictions of our theoretical model with the results of high-resolution N-body simulations.