Sub-quadratic (1+\eps)-approximate Euclidean Spanners, with Applications
Alexandr Andoni, Hengjie Zhang

TL;DR
This paper investigates the limits and possibilities of constructing efficient Euclidean graph spanners in high-dimensional spaces, introducing new subquadratic size constructions with applications to approximating Earth-Mover Distance.
Contribution
It proves the impossibility of certain spanners with stretch less than √2 and subquadratic size, and presents novel constructions of (1+ε)-approximate spanners with subquadratic size using extra nodes.
Findings
Spanners with stretch <√2 and subquadratic size are impossible.
Constructed (1+ε)-approximate spanners of size n^{2−Ω(ε^3)} and n^{2−Ω(ε^2)}.
Applied directed spanners to achieve faster approximation algorithms for Earth-Mover Distance.
Abstract
We study graph spanners for point-set in the high-dimensional Euclidean space. On the one hand, we prove that spanners with stretch <\sqrt{2} and subquadratic size are not possible, even if we add Steiner points. On the other hand, if we add extra nodes to the graph (non-metric Steiner points), then we can obtain (1+\eps)-approximate spanners of subquadratic size. We show how to construct a spanner of size n^{2-\Omega(\eps^3)}, as well as a directed version of the spanner of size n^{2-\Omega(\eps^2)}. We use our directed spanner to obtain an algorithm for computing (1+\eps)-approximation to Earth-Mover Distance (optimal transport) between two sets of size n in time n^{2-\Omega(\eps^2)}.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsComputational Geometry and Mesh Generation
