Push forward measures and concentration phenomena

Abstract: In this note we study how a concentration phenomenon can be transmitted from one measure

m u

to a push-forward measure

u

. In the first part, we push forward

m u

by

p i : s u p p (m u) i g h t a r r o w R e n

, where

p i x = f r a c x o r m x_{L} o r m x_{K}

, and obtain a concentration inequality in terms of the medians of the given norms (with respect to

m u

) and the Banach-Mazur distance between them. This approach is finer than simply bounding the concentration of the push forward measure in terms of the Banach-Mazur distance between

K

and

L

. As a corollary we show that any normed probability space with good concentration is far from any high dimensional subspace of the cube. In the second part, two measures

m u

and

u

are given, both related to the norm

o r m {c d o t}_{L}

, obtaining a concentration inequality in which it is involved the Banach-Mazur distance between

K

and

L

and the Lipschitz constant of the map that pushes forward

m u

into

u

. As an application, we obtain a concentration inequality for the cross polytope with respect to the normalized Lebesgue measure and the

e l l_{1}

norm.

Recommendations

Cites work

Cited in

(1)

Neural tangent kernel analysis to probe convergence in physics-informed neural solvers: PIKANs vs. PINNs

This page was built for publication: Push forward measures and concentration phenomena

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5417576)