English

Stabilizing isomorphisms from $\ell_p(\ell_2)$ into $L_p[0,1]$

Functional Analysis 2011-03-02 v1

Abstract

Let 1<p2<1<p\not=2<\infty, ϵ>0\epsilon>0 and let T:p(2)intoLp[0,1]T:\ell_p(\ell_2)\overset{into}{\rightarrow}L_p[0,1] be an isomorphism. Then there is a subspace Yp(2)Y\subset \ell_p(\ell_2) (1+ϵ)(1+\epsilon)-isomorphic to p(2)\ell_p(\ell_2) such that: TYT_{|Y} is an (1+ϵ)(1+\epsilon)-isomorphism and T(Y)T(Y) is KpK_p-complemented in Lp[0,1]L_p[0,1], with KpK_p depending only on pp. Moreover, Kp(1+ϵ)γpK_p\le (1+\epsilon)\gamma_p if p>2p>2 and Kp(1+ϵ)γp/(p1)K_p\le (1+\epsilon)\gamma_{p/(p-1)} if 1<p<21<p<2, where γr\gamma_r is the LrL_r norm of a standard Gaussian variable.

Keywords

Cite

@article{arxiv.1103.0047,
  title  = {Stabilizing isomorphisms from $\ell_p(\ell_2)$ into $L_p[0,1]$},
  author = {Ran Levy and Gideon Schechtman},
  journal= {arXiv preprint arXiv:1103.0047},
  year   = {2011}
}
R2 v1 2026-06-21T17:33:16.205Z