Presented By: Department of Mathematics
Colloquium Series Seminar
Singularity of random Bernoulli matrices
Let X_1,X_2,...,X_n be n mutually independent random vectors where each vector is uniformly distributed on vertices of the standard cube [-1,1]^n. What is the probability that the vectors are linearly dependent? Starting from a pioneering work of Komlos in the 1960-es, the question has attracted considerable attention. It has been conjectured that the probability is of the order (1/2+o(1))^n. Until recently, the best known upper estimate on the probability was (1/sqrt{2}+o(1))^n, due to Bourgain, Vu and Wood.
In this talk, I will discuss a positive resolution of the conjecture, as well as some further developments. Speaker(s): Konstantin Tikhomirov (Georgia Tech)
In this talk, I will discuss a positive resolution of the conjecture, as well as some further developments. Speaker(s): Konstantin Tikhomirov (Georgia Tech)
Co-Sponsored By
Explore Similar Events
-
Loading Similar Events...