On Oversquashing in Graph Neural Networks Through the Lens of Dynamical Systems

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Gravina, Alessio, Eliasof, Moshe, Gallicchio, Claudio, Bacciu, Davide, Schönlieb, Carola-Bibiane
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866912251140636672
author Gravina, Alessio
Eliasof, Moshe
Gallicchio, Claudio
Bacciu, Davide
Schönlieb, Carola-Bibiane
author_facet Gravina, Alessio
Eliasof, Moshe
Gallicchio, Claudio
Bacciu, Davide
Schönlieb, Carola-Bibiane
contents A common problem in Message-Passing Neural Networks is oversquashing -- the limited ability to facilitate effective information flow between distant nodes. Oversquashing is attributed to the exponential decay in information transmission as node distances increase. This paper introduces a novel perspective to address oversquashing, leveraging dynamical systems properties of global and local non-dissipativity, that enable the maintenance of a constant information flow rate. We present SWAN, a uniquely parameterized GNN model with antisymmetry both in space and weight domains, as a means to obtain non-dissipativity. Our theoretical analysis asserts that by implementing these properties, SWAN offers an enhanced ability to transmit information over extended distances. Empirical evaluations on synthetic and real-world benchmarks that emphasize long-range interactions validate the theoretical understanding of SWAN, and its ability to mitigate oversquashing.
format Preprint
id arxiv_https___arxiv_org_abs_2405_01009
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle On Oversquashing in Graph Neural Networks Through the Lens of Dynamical Systems
Gravina, Alessio
Eliasof, Moshe
Gallicchio, Claudio
Bacciu, Davide
Schönlieb, Carola-Bibiane
Machine Learning
A common problem in Message-Passing Neural Networks is oversquashing -- the limited ability to facilitate effective information flow between distant nodes. Oversquashing is attributed to the exponential decay in information transmission as node distances increase. This paper introduces a novel perspective to address oversquashing, leveraging dynamical systems properties of global and local non-dissipativity, that enable the maintenance of a constant information flow rate. We present SWAN, a uniquely parameterized GNN model with antisymmetry both in space and weight domains, as a means to obtain non-dissipativity. Our theoretical analysis asserts that by implementing these properties, SWAN offers an enhanced ability to transmit information over extended distances. Empirical evaluations on synthetic and real-world benchmarks that emphasize long-range interactions validate the theoretical understanding of SWAN, and its ability to mitigate oversquashing.
title On Oversquashing in Graph Neural Networks Through the Lens of Dynamical Systems
topic Machine Learning
url https://arxiv.org/abs/2405.01009