HEAT: Hyperbolic Embedding of Attributed Networks

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

Authors

Colleges, School and Institutes

Abstract

Finding a low dimensional representation of hierarchical, structured data described by a network remains a challenging problem in the machine learning community. An emerging approach is embedding networks into hyperbolic space since it can naturally represent a network’s hierarchical structure. However, existing hyperbolic embedding approaches cannot deal with attributed networks, where nodes are annotated with additional attributes. To overcome this, we propose HEAT (for Hyperbolic Embedding of Attributed Networks). HEAT first extracts training samples from the network that captures both topological and attribute node similarity and then learns a low-dimensional hyperboloid embedding using full Riemannian Stochastic Gradient Descent. We show that HEAT can outperform other network embedding algorithms on several common downstream tasks. As a general network embedding method, HEAT opens the door to hyperbolic manifold learning on a wide range of both attributed and unattributed networks.

Bibliographic note

Publisher Copyright: © 2020, Springer Nature Switzerland AG. Copyright: Copyright 2020 Elsevier B.V., All rights reserved.

Details

Original languageEnglish
Title of host publicationIntelligent Data Engineering and Automated Learning – IDEAL 2020 - 21st International Conference, 2020, Proceedings
EditorsCesar Analide, Paulo Novais, David Camacho, Hujun Yin
Publication statusPublished - 2020
Event21th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2020 - Guimaraes, Portugal
Duration: 4 Nov 20206 Nov 2020

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12489 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference21th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2020
Country/TerritoryPortugal
CityGuimaraes
Period4/11/206/11/20

Keywords

  • Hyperbolic embedding, Network embedding, Random walk