BIROn - Birkbeck Institutional Research Online

    On the hardness of computing the edit distance of shallow trees

    Charalampopoulos, Panagiotis and Gawrychowski, P. and Mozes, S. and Weimann, O. (2022) On the hardness of computing the edit distance of shallow trees. Lecture Notes in Computer Science 13617 , pp. 290-302. ISSN 0302-9743.

    [img]
    Preview
    Text
    camera_ready.pdf - Author's Accepted Manuscript

    Download (602kB) | Preview

    Abstract

    We consider the edit distance problem on rooted ordered trees parameterized by the trees’ depth. For two trees of size at most n and depth at most d, the state-of-the-art solutions of Zhang and Shasha [SICOMP 1989] and Demaine et al. [TALG 2009] have runtimes O(n^2 · d^2) and O(n^3), respectively, and are based on so-called decomposition algorithms. It has been recently shown by Bringmann et al. [TALG 2020] that, when d=Θ(n), one cannot compute the edit distance of two trees in O(n^{3−ϵ}) time (for any constant ϵ>0) under the APSP hypothesis. However, for small values of d, it is not known whether the O(n^2 · d^2) upper bound of Zhang and Shasha is optimal. We make the following twofold contribution. First, we show that under the APSP hypothesis there is no algorithm with runtime O(n^2 · d^{1−ϵ}) (for any constant ϵ>0) when d=poly(n). Second, we show that there is no decomposition algorithm that runs in time o(min{n^2 · d^2,n^3}).

    Metadata

    Item Type: Article
    Additional Information: International Symposium on String Processing and Information Retrieval. SPIRE 2022: String Processing and Information Retrieval, Proceedings. ISBN: 9783031206429
    School: Birkbeck Faculties and Schools > Faculty of Science > School of Computing and Mathematical Sciences
    Depositing User: Panagiotis Charalampopoulos
    Date Deposited: 30 Nov 2022 06:38
    Last Modified: 31 Oct 2023 01:10
    URI: https://eprints.bbk.ac.uk/id/eprint/49968

    Statistics

    Activity Overview
    6 month trend
    17Downloads
    6 month trend
    50Hits

    Additional statistics are available via IRStats2.

    Archive Staff Only (login required)

    Edit/View Item Edit/View Item