Discrete Analysis and Quantitative Geometry

forthcoming

This project lies at the crossroads of analysis, geometry and probability theory. DiAnQuGe aims to uncover quantitative properties of geometric objects using techniques from functional analysis, harmonic analysis and probability. Specifically, it is dedicated to the following directions: How well does a given metric space embed in a Banach space or a space with curvature bounds? In Section 1, I propose to investigate the (in)compatibility of the geometries of finite metrics, including expander-like objects and balls of Carnot groups, with Alexandrov spaces of bounded curvature. Based on a well-established analogy with nonlinear Banach space theory, this will require the development of suitable metric invariants, thus forming a link with discrete analysis. On a related direction, I propose a systematic way to develop such invariants for normed spaces, which will serve as a categorical implementation of Bourgain's proposed Ribe program. How does the spectrum of a voting system correlate with the behavior of individual voters? Voting systems are mathematically modeled by functions defined on the vertices of a hypercube in Euclidean space. In Section 2, I propose to utilize input from Fourier analysis, approximation theory and stochastic analysis to further understand the interaction of the spectrum and influences of Boolean functions, along with applications in computer science. How does a high-dimensional convex set look like? In Section 3, I propose to investigate sharp behaviors of isoperimetric profiles of convex sets using a mixture of geometric and probabilistic techniques, with a novel component from calculus of variations. Despite the different origins of these questions, all aforementioned objects share the common feature of having a large-yet-finite dimension, in an appropriate sense. This joint characteristic allows for the transfer of intuition and techniques from one field to the other and has been the driving force behind many recent advances.