Shahar Mendelson
(Technion)
"Geometric aspects of Learning Theory"
Abstract:
One of the main problems in Learning Theory is to estimate an unknown function by a function from a given family, using random data. It turns out that this question has very strong connections to Asymptotic Geometric Analysis, and in this talk I will present some of these connections. As an example, I will present a solution to the well-known reconstruction problem (studied, for example, by Donoho, Candes and Tao in recent years).