Sign is Not a Remedy: Multiset-to-Multiset Message Passing for Learning on Heterophilic Graphs

Langzhang Liang, Sunwoo Kim, Kijung Shin, Zenglin Xu, Shirui Pan, Yuan Qi
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:29621-29643, 2024.

Abstract

Graph Neural Networks (GNNs) have gained significant attention as a powerful modeling and inference method, especially for homophilic graph-structured data. To empower GNNs in heterophilic graphs, where adjacent nodes exhibit dissimilar labels or features, Signed Message Passing (SMP) has been widely adopted. However, there is a lack of theoretical and empirical analysis regarding the limitations of SMP. In this work, we unveil the potential pitfalls of SMP and their remedies. We first identify two limitations of SMP: undesirable representation update for multi-hop neighbors and vulnerability against oversmoothing issues. To overcome these challenges, we propose a novel message-passing function called Multiset to Multiset GNN (M2M-GNN). Our theoretical analyses and extensive experiments demonstrate that M2M-GNN effectively alleviates the limitations of SMP, yielding superior performance in comparison.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-liang24c, title = {Sign is Not a Remedy: Multiset-to-Multiset Message Passing for Learning on Heterophilic Graphs}, author = {Liang, Langzhang and Kim, Sunwoo and Shin, Kijung and Xu, Zenglin and Pan, Shirui and Qi, Yuan}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {29621--29643}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/liang24c/liang24c.pdf}, url = {https://proceedings.mlr.press/v235/liang24c.html}, abstract = {Graph Neural Networks (GNNs) have gained significant attention as a powerful modeling and inference method, especially for homophilic graph-structured data. To empower GNNs in heterophilic graphs, where adjacent nodes exhibit dissimilar labels or features, Signed Message Passing (SMP) has been widely adopted. However, there is a lack of theoretical and empirical analysis regarding the limitations of SMP. In this work, we unveil the potential pitfalls of SMP and their remedies. We first identify two limitations of SMP: undesirable representation update for multi-hop neighbors and vulnerability against oversmoothing issues. To overcome these challenges, we propose a novel message-passing function called Multiset to Multiset GNN (M2M-GNN). Our theoretical analyses and extensive experiments demonstrate that M2M-GNN effectively alleviates the limitations of SMP, yielding superior performance in comparison.} }
Endnote
%0 Conference Paper %T Sign is Not a Remedy: Multiset-to-Multiset Message Passing for Learning on Heterophilic Graphs %A Langzhang Liang %A Sunwoo Kim %A Kijung Shin %A Zenglin Xu %A Shirui Pan %A Yuan Qi %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-liang24c %I PMLR %P 29621--29643 %U https://proceedings.mlr.press/v235/liang24c.html %V 235 %X Graph Neural Networks (GNNs) have gained significant attention as a powerful modeling and inference method, especially for homophilic graph-structured data. To empower GNNs in heterophilic graphs, where adjacent nodes exhibit dissimilar labels or features, Signed Message Passing (SMP) has been widely adopted. However, there is a lack of theoretical and empirical analysis regarding the limitations of SMP. In this work, we unveil the potential pitfalls of SMP and their remedies. We first identify two limitations of SMP: undesirable representation update for multi-hop neighbors and vulnerability against oversmoothing issues. To overcome these challenges, we propose a novel message-passing function called Multiset to Multiset GNN (M2M-GNN). Our theoretical analyses and extensive experiments demonstrate that M2M-GNN effectively alleviates the limitations of SMP, yielding superior performance in comparison.
APA
Liang, L., Kim, S., Shin, K., Xu, Z., Pan, S. & Qi, Y.. (2024). Sign is Not a Remedy: Multiset-to-Multiset Message Passing for Learning on Heterophilic Graphs. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:29621-29643 Available from https://proceedings.mlr.press/v235/liang24c.html.

Related Material