Saved in:
Bibliographic Details
Main Authors: Karabulut, Tuğrul Hasan, Baytaş, İnci M.
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.00578
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910859066867712
author Karabulut, Tuğrul Hasan
Baytaş, İnci M.
author_facet Karabulut, Tuğrul Hasan
Baytaş, İnci M.
contents Graph Neural Networks (GNNs) set the state-of-the-art in representation learning for graph-structured data. They are used in many domains, from online social networks to complex molecules. Most GNNs leverage the message-passing paradigm and achieve strong performances on various tasks. However, the message-passing mechanism used in most models suffers from over-smoothing as a GNN's depth increases. The over-smoothing degrades GNN's performance due to the increased similarity between the representations of unrelated nodes. This study proposes an adaptive channel-wise message-passing approach to alleviate the over-smoothing. The proposed model, Channel-Attentive GNN, learns how to attend to neighboring nodes and their feature channels. Thus, much diverse information can be transferred between nodes during message-passing. Experiments with widely used benchmark datasets show that the proposed model is more resistant to over-smoothing than baselines and achieves state-of-the-art performances for various graphs with strong heterophily. Our code is at https://github.com/ALLab-Boun/CHAT-GNN.
format Preprint
id arxiv_https___arxiv_org_abs_2503_00578
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Channel-Attentive Graph Neural Networks
Karabulut, Tuğrul Hasan
Baytaş, İnci M.
Machine Learning
Graph Neural Networks (GNNs) set the state-of-the-art in representation learning for graph-structured data. They are used in many domains, from online social networks to complex molecules. Most GNNs leverage the message-passing paradigm and achieve strong performances on various tasks. However, the message-passing mechanism used in most models suffers from over-smoothing as a GNN's depth increases. The over-smoothing degrades GNN's performance due to the increased similarity between the representations of unrelated nodes. This study proposes an adaptive channel-wise message-passing approach to alleviate the over-smoothing. The proposed model, Channel-Attentive GNN, learns how to attend to neighboring nodes and their feature channels. Thus, much diverse information can be transferred between nodes during message-passing. Experiments with widely used benchmark datasets show that the proposed model is more resistant to over-smoothing than baselines and achieves state-of-the-art performances for various graphs with strong heterophily. Our code is at https://github.com/ALLab-Boun/CHAT-GNN.
title Channel-Attentive Graph Neural Networks
topic Machine Learning
url https://arxiv.org/abs/2503.00578