Learning on Large Graphs using Intersecting Communities

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Finkelshtein, Ben, Ceylan, İsmail İlkan, Bronstein, Michael, Levie, Ron
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910759816003584
author Finkelshtein, Ben
Ceylan, İsmail İlkan
Bronstein, Michael
Levie, Ron
author_facet Finkelshtein, Ben
Ceylan, İsmail İlkan
Bronstein, Michael
Levie, Ron
contents Message Passing Neural Networks (MPNNs) are a staple of graph machine learning. MPNNs iteratively update each node's representation in an input graph by aggregating messages from the node's neighbors, which necessitates a memory complexity of the order of the number of graph edges. This complexity might quickly become prohibitive for large graphs provided they are not very sparse. In this paper, we propose a novel approach to alleviate this problem by approximating the input graph as an intersecting community graph (ICG) -- a combination of intersecting cliques. The key insight is that the number of communities required to approximate a graph does not depend on the graph size. We develop a new constructive version of the Weak Graph Regularity Lemma to efficiently construct an approximating ICG for any input graph. We then devise an efficient graph learning algorithm operating directly on ICG in linear memory and time with respect to the number of nodes (rather than edges). This offers a new and fundamentally different pipeline for learning on very large non-sparse graphs, whose applicability is demonstrated empirically on node classification tasks and spatio-temporal data processing.
format Preprint
id arxiv_https___arxiv_org_abs_2405_20724
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Learning on Large Graphs using Intersecting Communities
Finkelshtein, Ben
Ceylan, İsmail İlkan
Bronstein, Michael
Levie, Ron
Machine Learning
Social and Information Networks
Message Passing Neural Networks (MPNNs) are a staple of graph machine learning. MPNNs iteratively update each node's representation in an input graph by aggregating messages from the node's neighbors, which necessitates a memory complexity of the order of the number of graph edges. This complexity might quickly become prohibitive for large graphs provided they are not very sparse. In this paper, we propose a novel approach to alleviate this problem by approximating the input graph as an intersecting community graph (ICG) -- a combination of intersecting cliques. The key insight is that the number of communities required to approximate a graph does not depend on the graph size. We develop a new constructive version of the Weak Graph Regularity Lemma to efficiently construct an approximating ICG for any input graph. We then devise an efficient graph learning algorithm operating directly on ICG in linear memory and time with respect to the number of nodes (rather than edges). This offers a new and fundamentally different pipeline for learning on very large non-sparse graphs, whose applicability is demonstrated empirically on node classification tasks and spatio-temporal data processing.
title Learning on Large Graphs using Intersecting Communities
topic Machine Learning
Social and Information Networks
url https://arxiv.org/abs/2405.20724