Average embedding conjecture for finite metric spaces into iterated llpll_p spaces

Let (M,dM)(\mathcal{M},d_\mathcal{M}) be an nn-point metric space. For p,D1p,D\geqslant 1, a bi-Lipschitz distortion-DD embedding of M\mathcal{M} into a metric space means an embedding whose Lipschitz distortion is at most DD.

There is a universal constant C1\mathsf{C}\geqslant 1 such that, for every p,D1p,D\geqslant 1, there exists n0=n0(p,D)Nn_0=n_0(p,D)\in\mathbb{N} with the following property: whenever nn0n\geqslant n_0, every nn-point metric space (M,dM)(\mathcal{M},d_\mathcal{M}) admits a bi-Lipschitz distortion-DD embedding into p(X)\ell_p(X) for some kk-dimensional normed space (X,X)(X,\|\cdot\|_X), where knC/(pD)k\leqslant n^{\mathsf{C}/(pD)}. Conceivably, this holds even with X=kX=\ell_\infty^k.

This conjecture predicts sharp dependence on pp in the preceding average John theorem and could have algorithmic applications; the stronger assertion with X=kX=\ell_\infty^k is also left open.

Sources & referencesView supporting material

Primary source

Assaf Naor, “An average John theorem”, arXiv:1905.01280 (2020).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.