Enregistré dans:
Détails bibliographiques
Auteurs principaux: Sulser, Aurelio, Wenckstern, Johann, Kuempel, Clara
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:https://arxiv.org/abs/2407.14387
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866911962173014016
author Sulser, Aurelio
Wenckstern, Johann
Kuempel, Clara
author_facet Sulser, Aurelio
Wenckstern, Johann
Kuempel, Clara
contents We propose GLAudio: Graph Learning on Audio representation of the node features and the connectivity structure. This novel architecture propagates the node features through the graph network according to the discrete wave equation and then employs a sequence learning architecture to learn the target node function from the audio wave signal. This leads to a new paradigm of learning on graph-structured data, in which information propagation and information processing are separated into two distinct steps. We theoretically characterize the expressivity of our model, introducing the notion of the receptive field of a vertex, and investigate our model's susceptibility to over-smoothing and over-squashing both theoretically as well as experimentally on various graph datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2407_14387
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle GLAudio Listens to the Sound of the Graph
Sulser, Aurelio
Wenckstern, Johann
Kuempel, Clara
Machine Learning
Artificial Intelligence
We propose GLAudio: Graph Learning on Audio representation of the node features and the connectivity structure. This novel architecture propagates the node features through the graph network according to the discrete wave equation and then employs a sequence learning architecture to learn the target node function from the audio wave signal. This leads to a new paradigm of learning on graph-structured data, in which information propagation and information processing are separated into two distinct steps. We theoretically characterize the expressivity of our model, introducing the notion of the receptive field of a vertex, and investigate our model's susceptibility to over-smoothing and over-squashing both theoretically as well as experimentally on various graph datasets.
title GLAudio Listens to the Sound of the Graph
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2407.14387