Volumes in High Dimension

Volumes in High Dimension
נפחים בממד גבוה

Spring semester 2018
March 15, 2018 - July 6, 2018

• Lecturers
  Bo'az Klartag and Uri Grupel 
   
• Classes
  Uri: Wednesday, 14:15 - 15:00, Ziskind 155
  Bo'az: Thursday, 10:45 - 11:30, Ziskind 1 and 11:45 - 12:30, Ziskind 155
   
• Course Description

  A priori, one would expect geometry in high-dimensional spaces - even in Euclidean spaces - to be rather complicated. Our experience in two and three dimensions seems to indicate that as the number of dimensions increases, the number of possible configurations grows rapidly, and we enter the realm of enormous, unimaginable diversity.

  Nevertheless, in this class we will see that dimensionality, when correctly viewed, may become a blessing. There are motifs in high-dimensional geometry, most notably the concentration of measure, which seem to compensate for the vast amount of different possibilities. Convexity is one of the ways in which to harness this and other motifs and thereby formulate clean, non-trivial theorems. 

  We plan to cover most of the following syllabus.

  Part I: High dimension 

  • Estimates for the central limit theorem for i.i.d random variables.
  • The isoperimetric inequality on the sphere, concentration of measure.
  • Maximal volume ellipsoid (John) and Dvoretzky's theorem: Any high-dimensional convex body contains approximately-spherical sections.
  • Analytic techniques: Poincare and Log-Sobolev inequalities, martingales, curvature.
  • Thin shell theorem, Gaussian marginals with geometric assumptions on the random variables in place of independence.
     

  Part II: Convexity  

  • Brunn-Minkowski inequality, concentration for uniformly convex sets.
  • Volume-ratio and Kashin's theorem (approximately-spherical sections of almost full dimension).
  • Slepian's lemma and Sudakov's inequality
  • Low M*-estimate, Johnson-Lindenstrauss
  • Complex interpolation and K-convexity.
  • Milman ellipsoid, quotient of subspace, Santalo & reverse-Santalo, reverse Brunn-Minkowksi.
     

• Prerequisites
  Familiarity with undergraduate probability, real analysis (say, Lebesgue measure) and Hilbert spaces.
   
• Final grade
  It will be based on the solution of the homework exercises. 
   
• Related literature

  There are a few books covering this material, here are some recommended ones:

  • Artstein-Avidan, Giannopoulos, Milman, Asymptotic Geometric Analysis, part I, 2015.

  • Aubrun, Szarek, Alice and Bob meet Banach, 2017.

  • Boucheron, Lugosi, Massart, Concentration inequalities, 2013.

  • Brazitikos, Giannopoulos, Valettas, Vritsiou, Geometry of Isotropic Convex Bodies, 2014.

  • Ledoux, The concentration of measure phenomenon, 2001.

  • Milman, Schechtman, Asymptotic theory of finite-dimensional normed spaces, 1986.

  • Pisier, The volume of convex bodies and Banach space geometry, 1989.

  • Vershynin, High-Dimensional Probability. The book hasn't appeared yet, a draft is available here.

• Comments 
  • An argument related to Slepian's lemma.
     
• Homework (written by Uri)
  • Exercise sheet no. 1
  • Exercise sheet no. 2
  • Exercise sheet no. 3
  • Exercise sheet no. 4
  • Exercise sheet no. 5
  • Exercise sheet no. 6
  • Exercise sheet no. 7
  • Exercise sheet no. 8
  • Exercise sheet no. 9
  • Remarks on student submissions so far

"Il serait paradoxical que le grand nombre des variables fût une cause de simplicité", P. Levy, Problemes concrets d'analyse fonctionnelle, page xi.

(Free translation: It is paradoxical that simplicity could arise from a large number of variables)