English

Lossless Prioritized Embeddings

Data Structures and Algorithms 2019-07-17 v1

Abstract

Given metric spaces (X,d)(X,d) and (Y,ρ)(Y,\rho) and an ordering x1,x2,,xnx_1,x_2,\ldots,x_n of (X,d)(X,d), an embedding f:XYf: X \rightarrow Y is said to have a prioritized distortion α()\alpha(\cdot), if for any pair xj,xx_j,x' of distinct points in XX, the distortion provided by ff for this pair is at most α(j)\alpha(j). If YY is a normed space, the embedding is said to have prioritized dimension β()\beta(\cdot), if f(xj)f(x_j) may have nonzero entries only in its first β(j)\beta(j) coordinates. The notion of prioritized embedding was introduced by \cite{EFN15}, where a general methodology for constructing such embeddings was developed. Though this methodology enables \cite{EFN15} to come up with many prioritized embeddings, it typically incurs some loss in the distortion. This loss is problematic for isometric embeddings. It is also troublesome for Matousek's embedding of general metrics into \ell_\infty, which for a parameter k=1,2,k = 1,2,\ldots, provides distortion 2k12k-1 and dimension O(klognn1/k)O(k \log n \cdot n^{1/k}). In this paper we devise two lossless prioritized embeddings. The first one is an isometric prioritized embedding of tree metrics into \ell_\infty with dimension O(logj)O(\log j). The second one is a prioritized Matousek's embedding of general metrics into \ell_\infty, which provides prioritized distortion 2klogjlogn12 \lceil k {{\log j} \over {\log n}} \rceil - 1 and dimension O(klognn1/k)O(k \log n \cdot n^{1/k}), again matching the worst-case guarantee 2k12k-1 in the distortion of the classical Matousek's embedding. We also provide a dimension-prioritized variant of Matousek's embedding. Finally, we devise prioritized embeddings of general metrics into (single) ultra-metric and of general graphs into (single) spanning tree with asymptotically optimal distortion.

Keywords

Cite

@article{arxiv.1907.06983,
  title  = {Lossless Prioritized Embeddings},
  author = {Michael Elkin and Ofer Neiman},
  journal= {arXiv preprint arXiv:1907.06983},
  year   = {2019}
}

Comments

abstract was shortened

R2 v1 2026-06-23T10:22:07.978Z