Speaker: Dan Shiber
Affiliation: UCLA
Title: Information Theory and Large Random Matrices.
Time and Place: Thursday, April 19, 3:00-3:50pm, Milner 317.
Abstract: In the first part of this talk we review classical information theory (following Amari and Nagaoka) including Gaussian and exponential families, Fisher information measure and metric, the Cramer-Rao theorem and Legendre transform of pressure. In the second part we construct the analogous quantities for large random matrices and relate our construction to free probability.