In this talk, we consider log|det(M - s)| of random matrix M from the Laguerre \beta ensemble in the case s is near the upper edge of the Marchenko-Pastur law, and discuss its convergence to a Gaussian random variable as the size of of the matrix grows. We first introduce the Laguerre ensemble for general \beta>0, then examine the log determinant from the perspective of linear spectral statistics. Time permitting, we will discuss the starting point of the CLT analysis, and the connection to a spherical spin glass model.

The talk is based on joint work with Elizabeth Collins-Woodfin. Speaker(s): Han Le (University of Michigan)

The talk is based on joint work with Elizabeth Collins-Woodfin. Speaker(s): Han Le (University of Michigan)

Building: | East Hall |
---|---|

Event Type: | Workshop / Seminar |

Tags: | Mathematics |

Source: | Happening @ Michigan from Department of Mathematics |