Scalable Message Passing Neural Networks: No Need for Attention in Large Graph Representation Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Borde, Haitz Sáez de Ocáriz, Lukoianov, Artem, Kratsios, Anastasis, Bronstein, Michael, Dong, Xiaowen
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910046665834496
author Borde, Haitz Sáez de Ocáriz
Lukoianov, Artem
Kratsios, Anastasis
Bronstein, Michael
Dong, Xiaowen
author_facet Borde, Haitz Sáez de Ocáriz
Lukoianov, Artem
Kratsios, Anastasis
Bronstein, Michael
Dong, Xiaowen
contents We propose Scalable Message Passing Neural Networks (SMPNNs) and demonstrate that, by integrating standard convolutional message passing into a Pre-Layer Normalization Transformer-style block instead of attention, we can produce high-performing deep message-passing-based Graph Neural Networks (GNNs). This modification yields results competitive with the state-of-the-art in large graph transductive learning, particularly outperforming the best Graph Transformers in the literature, without requiring the otherwise computationally and memory-expensive attention mechanism. Our architecture not only scales to large graphs but also makes it possible to construct deep message-passing networks, unlike simple GNNs, which have traditionally been constrained to shallow architectures due to oversmoothing. Moreover, we provide a new theoretical analysis of oversmoothing based on universal approximation which we use to motivate SMPNNs. We show that in the context of graph convolutions, residual connections are necessary for maintaining the universal approximation properties of downstream learners and that removing them can lead to a loss of universality.
format Preprint
id arxiv_https___arxiv_org_abs_2411_00835
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Scalable Message Passing Neural Networks: No Need for Attention in Large Graph Representation Learning
Borde, Haitz Sáez de Ocáriz
Lukoianov, Artem
Kratsios, Anastasis
Bronstein, Michael
Dong, Xiaowen
Machine Learning
We propose Scalable Message Passing Neural Networks (SMPNNs) and demonstrate that, by integrating standard convolutional message passing into a Pre-Layer Normalization Transformer-style block instead of attention, we can produce high-performing deep message-passing-based Graph Neural Networks (GNNs). This modification yields results competitive with the state-of-the-art in large graph transductive learning, particularly outperforming the best Graph Transformers in the literature, without requiring the otherwise computationally and memory-expensive attention mechanism. Our architecture not only scales to large graphs but also makes it possible to construct deep message-passing networks, unlike simple GNNs, which have traditionally been constrained to shallow architectures due to oversmoothing. Moreover, we provide a new theoretical analysis of oversmoothing based on universal approximation which we use to motivate SMPNNs. We show that in the context of graph convolutions, residual connections are necessary for maintaining the universal approximation properties of downstream learners and that removing them can lead to a loss of universality.
title Scalable Message Passing Neural Networks: No Need for Attention in Large Graph Representation Learning
topic Machine Learning
url https://arxiv.org/abs/2411.00835