Quantitative affine approximation for UMD targets

Abstract: It is shown here that if

(Y, | c d o t |_{Y})

is a Banach space in which martingale differences are unconditional (a UMD Banach space) then there exists

c = c (Y) i n (0, i n f t y)

with the following property. For every

n i n m a t h b b N

and

v a r e p s i l o n i n (0, 1 / 2]

, if

(X, | c d o t |_{X})

is an

n

-dimensional normed space with unit ball

B_{X}

and

f : B_{X} o Y

is a

1

-Lipschitz function then there exists an affine mapping

L a m b d a : X o Y

and a sub-ball

B^{*} = y + h o B_{X} s u b s e t e q B_{X}

of radius

h o g e e x p (- (1 / v a r e p s i l o n)^{c n})

such that

| f (x) - L a m b d a (x) |_{Y} l e v a r e p s i l o n h o

for all

x i n B^{*}

. This estimate on the macroscopic scale of affine approximability of vector-valued Lipschitz functions is an asymptotic improvement (as

n o i n f t y

) over the best previously known bound even when

X

is

m a t h b b R^{n}

equipped with the Euclidean norm and

Y

is a Hilbert space.

Recommendations

Cites work

Cited in

(8)

This page was built for publication: Quantitative affine approximation for UMD targets