Computing Optimal Assignments in Linear Time for Approximate Graph Matching

Nils Kriege, Pierre-Louis Giscard, Franka Bause, Richard Charles Wilson

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Finding an optimal assignment between two sets of
objects is a fundamental problem arising in many applications,
including the matching of ‘bag-of-words’ representations in
natural language processing and computer vision. Solving the
assignment problem typically requires cubic time and its pairwise
computation is expensive on large datasets. In this paper, we
develop an algorithm which can find an optimal assignment in
linear time when the cost function between objects is represented
by a tree distance. We employ the method to approximate the edit
distance between two graphs by matching their vertices in linear
time. To this end, we propose two tree distances, the first of which
reflects discrete and structural differences between vertices, and
the second of which can be used to compare continuous labels.
We verify the effectiveness and efficiency of our methods using
synthetic and real-world datasets.
Original languageEnglish
Title of host publicationInternational Conference on Data Mining
PublisherIEEE Computer Society
Number of pages10
Publication statusAccepted/In press - 6 Sep 2019

Cite this