On the Komlos Conjecture, or, how to Control your Brownian Motion
The Komlos Conjecture is a notorious open problem in convex geometry and discrepancy theory.
In this talk, starting with the history and motivation for the problem and discussing some previous works,
I will describe our recent progress on the problem. Our result is based on various ingredients such as Gaussian processes,
operator theory, convex geometry and semidefinite programming duality and the talk will aim to give a flavor of these
areas and how they come together.
Based on joint work with Daniel Dadush (CWI) and Shashwat Garg (TU/e).