GraphAlign: Pretraining One Graph Neural Network on Multiple Graphs via Feature Alignment

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Hou, Zhenyu, Li, Haozhan, Cen, Yukuo, Tang, Jie, Dong, Yuxiao
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866929374702338048
author Hou, Zhenyu
Li, Haozhan
Cen, Yukuo
Tang, Jie
Dong, Yuxiao
author_facet Hou, Zhenyu
Li, Haozhan
Cen, Yukuo
Tang, Jie
Dong, Yuxiao
contents Graph self-supervised learning (SSL) holds considerable promise for mining and learning with graph-structured data. Yet, a significant challenge in graph SSL lies in the feature discrepancy among graphs across different domains. In this work, we aim to pretrain one graph neural network (GNN) on a varied collection of graphs endowed with rich node features and subsequently apply the pretrained GNN to unseen graphs. We present a general GraphAlign method that can be seamlessly integrated into the existing graph SSL framework. To align feature distributions across disparate graphs, GraphAlign designs alignment strategies of feature encoding, normalization, alongside a mixture-of-feature-expert module. Extensive experiments show that GraphAlign empowers existing graph SSL frameworks to pretrain a unified and powerful GNN across multiple graphs, showcasing performance superiority on both in-domain and out-of-domain graphs.
format Preprint
id arxiv_https___arxiv_org_abs_2406_02953
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle GraphAlign: Pretraining One Graph Neural Network on Multiple Graphs via Feature Alignment
Hou, Zhenyu
Li, Haozhan
Cen, Yukuo
Tang, Jie
Dong, Yuxiao
Machine Learning
Graph self-supervised learning (SSL) holds considerable promise for mining and learning with graph-structured data. Yet, a significant challenge in graph SSL lies in the feature discrepancy among graphs across different domains. In this work, we aim to pretrain one graph neural network (GNN) on a varied collection of graphs endowed with rich node features and subsequently apply the pretrained GNN to unseen graphs. We present a general GraphAlign method that can be seamlessly integrated into the existing graph SSL framework. To align feature distributions across disparate graphs, GraphAlign designs alignment strategies of feature encoding, normalization, alongside a mixture-of-feature-expert module. Extensive experiments show that GraphAlign empowers existing graph SSL frameworks to pretrain a unified and powerful GNN across multiple graphs, showcasing performance superiority on both in-domain and out-of-domain graphs.
title GraphAlign: Pretraining One Graph Neural Network on Multiple Graphs via Feature Alignment
topic Machine Learning
url https://arxiv.org/abs/2406.02953