Low distortion embeddings for edit distance

Rafail Ostrovsky*, Yuval Rabani

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

77 Scopus citations

Abstract

We show that {0, 1} d endowed with edit distance embeds into ℓ 1 with distortion 2 O(log d log log d). We further show efficient implementation of the embedding that yield solutions to various computational problems involving edit distance. These include sketching, communication complexity, nearest neighbor search. For all these problems, we improve upon previous bounds.

Original languageAmerican English
Article number23
JournalJournal of the ACM
Volume54
Issue number5
DOIs
StatePublished - 1 Oct 2007
Externally publishedYes

Keywords

  • Communication complexity
  • Computations on discrete structures
  • Dimension reduction
  • Edit distance
  • Levenshtein distance
  • Metric embeddings
  • Nearest neighbor search
  • Pattern matching
  • Sketching

Fingerprint

Dive into the research topics of 'Low distortion embeddings for edit distance'. Together they form a unique fingerprint.

Cite this