Objetivo
"The proposed project lies at the intersection of the local theory of Banach spaces (more precisely ""Asymptotic Geometric Analysis""), Global Riemannian Geometry, and the study of isoperimetric and concentration properties of such spaces. We will study Riemannian manifolds endowed with a probability measure, whose (generalized Ricci) curvature is non-negative (""convex""), or more generally, bounded below (""semi-convex""); an important example is that of the uniform measure on a convex bounded domain in Euclidean space. Despite the immense diversity of these structures, even in the Euclidean case, it is known that they exhibit various unifying geometric and probabilistic properties. In this project, we are interested in various quantitative manifestations of the concentration of measure on these spaces, as their dimension tends to infinity. These include isoperimetric inequalities, providing lower bounds on the boundary measure of sets; Sobolev-type inequalities, such as the classical Poincar\'e (or Spectral-Gap) and logarithmic-Sobolev inequalities; and concentration of measure of various Lipschitz functionals, such as the distance functional. All present conjectures suggest that despite the great diversity, convexity and high-dimensionality serve as unifying forces which render all of these spaces not very different from some canonical ones, like the uniform measure on a Euclidean ball or hyper-cube. In recent years there has been much progress in the analysis of these and related questions. The proposed project intends to deepen and extend our qualitative and quantitative understanding of isoperimetric and concentration inequalities on high-dimensional convex and semi-convex manifolds-with-density in general, and on log-concave measures and convex bodies in particular."
Ámbito científico
Convocatoria de propuestas
FP7-PEOPLE-2011-CIG
Consulte otros proyectos de esta convocatoria
Régimen de financiación
MC-CIG - Support for training and career development of researcher (CIG)Coordinador
32000 Haifa
Israel