Backhausz Ágnes (ELTE and Rényi): Graph limits and the eigenvectors of random sign matrices

The theory of random matrices and graph limits are both actively studied fields in probability theory and combinatorics. In this talk, our goal is to briefly present a possible connection of these two areas, in particular, a convergence notion coming from graph limits (action convergence) and its possible application for the eigenvectors of random sign matrices. We also present our results on the empirical distribution of the eigenvectors of these non-Hermitian random matrices, in which each entry has value $\pm 1/\sqrt n$, independently of each other. Joint work with Balázs Szegedy.