Abstract:It is well known that fluctuations of marginals of high-dimensional random vectors that satisfy a certain concentration estimate called the thin shell condition are approximately Gaussian. In this article we identify a general condition on a sequence of high-dimensional random vectors under which one can identify the exponential decay rate of large deviation probabilities of the corresponding sequence of marginals. More precisely, consider the projection of an n-dimensional random vector onto a random k n -dimensional basis, k n ≤ n , drawn uniformly from the Haar measure on the Stiefel manifold of orthonormal k n -frames in R n , in three different asymptotic regimes as n → ∞ : “constant” ( k n = k ), “sublinear” ( k n → ∞ but k n / n → 0 ) and “linear” ( k n / n → λ with 0 < λ ≤ 1 ). When the sequence of random vectors satisfies a certain “asymptotic thin shell condition”, we establish large deviation principles for the corresponding sequence of random projections in the constant regime, and for the sequence of empirical measures of the coordinates of the random projections in the sublinear and linear regimes. We also establish large deviation principles for scaled ℓ q norms of the random projections in all three regimes. Moreover, we show that the asymptotic thin shell condition holds for various sequences of random vectors of interest, including the uniform measure on suitably scaled ℓ p n balls, for p ∈ [ 1 , ∞ ) , and generalized Orlicz balls defined via a superquadratic function, as well as a class of Gibbs measures with superquadratic interaction potential. Along the way, we obtain logarithmic asymptotics of volumes of high-dimensional Orlicz balls, which may be of independent interest. We also show that the decay rate of large deviation probabilities of Euclidean norms of multi-dimensional projections of ℓ p n balls, when p ∈ [ 1 , 2 ) , exhibits an unexpected phase transition in the sublinear regime, thus disproving an earlier conjecture due to Alonso-Gutiérrez et al. Random projections of high-dimensional random vectors are of interest in a range of fields including asymptotic convex geometry and high-dimensional statistics.

Algorithms and a Library for the Exact Computation of the Cumulative Distribution Function of the Euclidean Distance Between a Point and a Random Variable Uniformly Distributed in Disks, Balls, or Polygones and Application to Probabilistic Seismic Hazard Analysis

Distance Distribution Between Two Random Nodes in Arbitrary Polygons

Modeling the Probability Distribution of Positional Errors Incurred by Residential Address Geocoding.

A Variational Approach to a Cumulative Distribution Function Estimation Problem Under Stochastic Ambiguity

Generating Probability Distributions on Intervals and Spheres: Convex Decomposition

Approximating the Earth Mover's Distance between sets of geometric objects

A Vecchia Approximation for High-Dimensional Gaussian Cumulative Distribution Functions Arising from Spatial Data

Random Distances Associated with Arbitrary Triangles: A Systematic Approach between Two Random Points

PDE characterisation of geometric distribution functions and quantiles

A generalization for the expected value of the earth mover's distance

Fluctuations of the connectivity threshold and largest nearest-neighbour link

Asymptotic expected $T$-functionals of random polytopes with applications to $L_p$ surface areas

Pointwise density estimation on metric spaces and applications in seismology

Testing Thresholds for High-Dimensional Sparse Random Geometric Graphs

Smooth Distance Approximation

Random approximation of convex bodies in Hausdorff metric

An asymptotic thin shell condition and large deviations for random multidimensional projections

Distribution of Test Statistic for Euclidean Distance Matrices

A distance theorem for inhomogenous random rectangular matrices