
Ponente:  Wolfgang Globke (Universität Wien, Vienna, Austria)  Lugar:  Salón G101  Hora:  16:00 hrs  Resumen:  Information geometry aims to understand the geometric structure of
certain problems in statistics and information theory with the methods of differential geometry. A parameterized family of probability distributions can be endowed with a Riemannian metric, the Fisher metric, which appears naturally in the context of parameter estimations. Some of these families have remarkable geometric properties, for instance, the family of univariate normal distributions with the Fisher metric is isometric to the hyperbolic plane.
In this talk, I provide a quick motivation for the terminology coming
from information theory and statistics, and then illustrate information
geometry by some examples and some interesting results. Time permitting,
I will talk about practical applications of information geometry. 
