On the hardness of learning queries from tree structured data

被引:0
|
作者
Xianmin Liu
Jianzhong Li
机构
[1] Harbin Institute of Technology,
来源
关键词
Tree Query Learning; Hardness; Inapproximability;
D O I
暂无
中图分类号
学科分类号
摘要
The problem of learning queries from tree structured data is studied by this paper. A tree structured data is modeled as a node-labeled tree T\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$T$$\end{document}, and applying a query q\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$q$$\end{document} on T\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$T$$\end{document} will return a set q(T)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$q(T)$$\end{document} which is a subset of nodes in T\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$T$$\end{document}. For a tree-node pair (T,t)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$(T,t)$$\end{document} where t\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t$$\end{document} is a node in T\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$T$$\end{document}, q\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$q$$\end{document} is called to accept the pair if t∈q(T)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t\in {q(T)}$$\end{document}, and reject the pair if t∉q(T)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$t\notin {q(T)}$$\end{document}. For some query class L\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal{L }$$\end{document}, given tree-node pair sets Ep\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_p$$\end{document} and En\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_n$$\end{document}, the tree query learning problem is to find a query q∈L\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$q\in \mathcal{L }$$\end{document} such that (1) q\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$q$$\end{document} rejects all pairs in En\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_n$$\end{document}, and (2) the size of pairs in Ep\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$E_p$$\end{document} accepted by q\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$q$$\end{document} is maximized. On four different query classes Q/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal Q ^{\tiny /}$$\end{document}, Q/,∗\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal Q ^{\tiny /,*}$$\end{document}, Q/,//\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal Q ^{\tiny /,//}$$\end{document} and Q/,[]\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal Q ^{\tiny /,[]}$$\end{document}, this paper studies the hardness of the corresponding tree query learning problems. For Q/\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal Q ^{\tiny /}$$\end{document}, a PTime algorithm is given. For Q/,∗\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal Q ^{\tiny /,*}$$\end{document} and Q/,//\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal Q ^{\tiny /,//}$$\end{document}, the NP-complete results are shown. For Q/,[]\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal Q ^{\tiny /,[]}$$\end{document}, the problem is shown to be NP-hard by considering two constrained fragments of Q/,[]\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal Q ^{\tiny /,[]}$$\end{document}. Also, for Q/,∗\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal Q ^{\tiny /,*}$$\end{document}, Q/,[]\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal Q ^{\tiny /,[]}$$\end{document} and Q/,//\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\mathcal Q ^{\tiny /,//}$$\end{document}, it is shown that there are no n1−ϵ\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$n^{1-\epsilon }$$\end{document}-approximation algorithms for any ϵ>0\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\epsilon >0$$\end{document}.
引用
收藏
页码:670 / 684
页数:14
相关论文
共 50 条
  • [21] Evolution of Multiple Tree Structured Patterns from Tree-Structured Data Using Clustering
    Nagamine, Masatoshi
    Miyahara, Tetsuhiro
    Kuboyama, Tetsuji
    Ueda, Hiroaki
    Takahashi, Kenichi
    [J]. AI 2008: ADVANCES IN ARTIFICIAL INTELLIGENCE, PROCEEDINGS, 2008, 5360 : 500 - +
  • [22] Tree2Vector: Learning a Vectorial Representation for Tree-Structured Data
    Zhang, Haijun
    Wang, Shuang
    Xu, Xiaofei
    Chow, Tommy W. S.
    Wu, Q. M. Jonathan
    [J]. IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2018, 29 (11) : 5304 - 5318
  • [23] Learning from data with structured missingness
    Robin Mitra
    Sarah F. McGough
    Tapabrata Chakraborti
    Chris Holmes
    Ryan Copping
    Niels Hagenbuch
    Stefanie Biedermann
    Jack Noonan
    Brieuc Lehmann
    Aditi Shenvi
    Xuan Vinh Doan
    David Leslie
    Ginestra Bianconi
    Ruben Sanchez-Garcia
    Alisha Davies
    Maxine Mackintosh
    Eleni-Rosalina Andrinopoulou
    Anahid Basiri
    Chris Harbron
    Ben D. MacArthur
    [J]. Nature Machine Intelligence, 2023, 5 : 13 - 23
  • [24] Learning from data with structured missingness
    Mitra, Robin
    McGough, Sarah F.
    Chakraborti, Tapabrata
    Holmes, Chris
    Copping, Ryan
    Hagenbuch, Niels
    Biedermann, Stefanie
    Noonan, Jack
    Lehmann, Brieuc
    Shenvi, Aditi
    Doan, Xuan Vinh
    Leslie, David
    Bianconi, Ginestra
    Sanchez-Garcia, Ruben
    Davies, Alisha
    Mackintosh, Maxine
    Andrinopoulou, Eleni-Rosalina
    Basiri, Anahid
    Harbron, Chris
    MacArthur, Ben D.
    [J]. NATURE MACHINE INTELLIGENCE, 2023, 5 (01) : 13 - 23
  • [25] Fuzzy Matching of Web Queries to Structured Data
    Cheng, Tao
    Lauw, Hady W.
    Paparizos, Stelios
    [J]. 26TH INTERNATIONAL CONFERENCE ON DATA ENGINEERING ICDE 2010, 2010, : 713 - 716
  • [26] Efficiently Answering Reachability Queries for Tree-Structured Data in Repetitive Prime Number Labeling Schemes
    Ahn, Jinhyun
    Lee, Taewhi
    Im, Dong-Hyuk
    [J]. APPLIED SCIENCES-BASEL, 2018, 8 (05):
  • [27] Decision tree construction from multidimensional structured data
    Watanuma, Tomoki
    Ozaki, Tomonobu
    Ohkawa, Takenao
    [J]. ICDM 2006: SIXTH IEEE INTERNATIONAL CONFERENCE ON DATA MINING, WORKSHOPS, 2006, : 237 - +
  • [28] Learning metrics between tree structured data: Application to image recognition
    Boyer, Laurent
    Habrard, Amaury
    Sebban, Marc
    [J]. MACHINE LEARNING: ECML 2007, PROCEEDINGS, 2007, 4701 : 54 - +
  • [29] Linear Online Learning over Structured Data with Distributed Tree Kernels
    Filice, Simone
    Croce, Danilo
    Basili, Roberto
    Zanzotto, Fabio Massimo
    [J]. 2013 12TH INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND APPLICATIONS (ICMLA 2013), VOL 1, 2013, : 123 - 128
  • [30] Hidden Tree Markov Networks: Deep and Wide Learning for Structured Data
    Bacciu, Davide
    [J]. 2017 IEEE SYMPOSIUM SERIES ON COMPUTATIONAL INTELLIGENCE (SSCI), 2017, : 3020 - 3027