Metric recovery from directed unweighted graphs
Name
Metric recovery.pdf
Size
2.73 MB
Format
Adobe PDF
Checksum (MD5)
843e086e6ffd395e95e8918884ccb173
Author(s)
Hashimoto, Tatsunori Benjamin
Sun, Yi
Jaakkola, Tommi S
Date Issued
May 2015
Journal
Proceedings of Machine Learning Research
Publisher
PMLR
Citation
Hashimoto, Tatsunori, Yi Sun, and Tommi S. Jaakkola. "Metric recovery from directed unweighted graphs." Proceedings of the 18th International Conference
on Artificial Intelligence and Statistics (AISTATS) 2015, 9-12 May, 2015, San Diego, California, PLMR, 2015. © The Authors
Version
Author's final manuscript
Abstract
We analyze directed, unweighted graphs obtained from x[subscript i] ∈ R[superscript d] by connecting vertex i to j iff |x[subscript i] − x[subscript j]| < ε(x[subscript i]). Examples of such graphs include k-nearest neighbor graphs, where ε(x[subscript i]) varies from point to point, and, arguably, many real-world graphs such as copurchasing graphs. We ask whether we can recover the underlying Euclidean metric ε(x[subscript i]) and the associated density p(xi) given only the directed graph and d. We show that consistent recovery is possible up to isometric scaling when the vertex degree is at least ω(n[superscript 2/(2+d)] log(n)[superscript d/(d+2)]). Our estimator is based on a careful characterization of a random walk over the directed graph and the associated continuum limit. As an algorithm, it resembles the PageRank centrality metric. We demonstrate empirically that the estimator performs well on simulated examples as well as on real-world co-purchasing graphs even with a small number of points and degree scaling as low as log(n).
MIT Department
Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
Massachusetts Institute of Technology. Department of Mathematics
Terms of Use
Creative Commons Attribution-Noncommercial-Share Alike
Persistent DSpace Link
DOI of Published Version
http://proceedings.mlr.press/v38/
