Maximal Independent Sets for Pooling in Graph Neural Networks - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Maximal Independent Sets for Pooling in Graph Neural Networks

Résumé

Convolutional Neural Networks (CNNs) have enabled major advances in image classification through convolution and pooling. In particular, image pooling transforms a connected discrete lattice into a reduced lattice with the same connectivity and allows reduction functions to consider all pixels in an image. However, there is no pooling that satisfies these properties for graphs. In fact, traditional graph pooling methods suffer from at least one of the following drawbacks: Graph disconnection or overconnection, low decimation ratio, and deletion of large parts of graphs. In this paper, we present three pooling methods based on the notion of maximal independent sets that avoid these pitfalls. Our experimental results confirm the relevance of maximal independent set constraints for graph pooling.
Fichier principal
Vignette du fichier
Maximal Independent Sets for Pooling in Graph Neural Networks.pdf (477.58 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04160860 , version 1 (22-07-2023)

Identifiants

Citer

Stevan Stanovic, Benoit Gaüzère, Luc Brun. Maximal Independent Sets for Pooling in Graph Neural Networks. 13th IAPR-TC15 International Workshop on Graph-Based Representations in Pattern Recognition (GbR 2023), Sep 2023, Vietri Sul Mare, Italy. ⟨10.1007/978-3-031-42795-4_11⟩. ⟨hal-04160860⟩
27 Consultations
50 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More