A simplified second-order Gaussian Poincaré inequality in discrete setting with applications (Q2686615)

In the current work, the authors provide a significantly simplified second-order Gaussian Poincaré inequality for normal approximation of functionals of infinitely many Rademacher random variables. Their approach depends on fewer terms than the existing approaches, and leads to simplifications and improvements of bounds on the Kolmogorov distance between quantities of interest in a variety of random models. More precisely, the main result can be applied in order to study the asymptotic normality of several quantities such as: infinite weighted 2-runs, subgraph counts and vertices of fixed degrees in Erdős-Rényi random graphs, isolated faces in the random \(\kappa\)-complex, vertices of fixed degree in hypercube percolation. Besides the applications to random graphs, the results are also interesting from a more theoretical point of view in the context of the fourth moment theorem. The proof of second-order Gaussian Poincaré inequality is based on a new bound for normal approximation of non-linear functionals of an infinite Rademacher sequence, and it is also of independent interest.