Τρίτη 6 Απριλίου, 11:15 – 12:00. Η ομιλία θα γίνει ONLINE.
Ομιλητής: Κωνσταντίνος Τύρος (ΕΚΠΑ)
Τίτλος: Concentration estimates for functions of finite high-dimensional random arrays
Περίληψη: The concentration of measure, roughly speaking, refers to the phenomenon that a function f, that depends smoothly on its variables, is essentially constant as long as the number of variables is large enough. Dropping the smoothness assumptions makes this phenomenon no longer valid. In 2016 in a joint work with P. Dodos and V. Kanellopoulos some form of concentration was obtained under a mild integrability condition for f(X), where X is a random vector with independent entries. In this talk, we shall present a result that extends the latter towards two directions. We substitute the random vector by a d-dimensional random array, where d is a positive integer, whose distribution is invariant under certain symmetries and we relax the independence condition on its entries. If time permits we will discuss some connections with combinatorics. This is a joint work with P. Dodos and P. Valettas.