Label-Wise Graph Convolutional Network for Heterophilic Graphs

Enyan Dai, Shijie Zhou, Zhimeng Guo, Suhang Wang
Proceedings of the First Learning on Graphs Conference, PMLR 198:26:1-26:21, 2022.

Abstract

Graph Neural Networks (GNNs) have achieved remarkable performance in modeling graphs for various applications. However, most existing GNNs assume the graphs exhibit strong homophily in node labels, i.e., nodes with similar labels are connected in the graphs. They fail to generalize to heterophilic graphs where linked nodes may have dissimilar labels and attributes. Therefore, in this paper, we investigate a novel framework that performs well on graphs with either homophily or heterophily. More specifically, we propose a label-wise message passing mechanism to avoid the negative effects caused by aggregating dissimilar node representations and preserve the heterophilic contexts for representation learning. We further propose a bi-level optimization method to automatically select the model for graphs with homophily/heterophily. Theoretical analysis and extensive experiments demonstrate the effectiveness of our proposed framework for node classification on both homophilic and heterophilic graphs.

Cite this Paper

BibTeX
@InProceedings{pmlr-v198-dai22b, title = {Label-Wise Graph Convolutional Network for Heterophilic Graphs}, author = {Dai, Enyan and Zhou, Shijie and Guo, Zhimeng and Wang, Suhang}, booktitle = {Proceedings of the First Learning on Graphs Conference}, pages = {26:1--26:21}, year = {2022}, editor = {Rieck, Bastian and Pascanu, Razvan}, volume = {198}, series = {Proceedings of Machine Learning Research}, month = {09--12 Dec}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v198/dai22b/dai22b.pdf}, url = {https://proceedings.mlr.press/v198/dai22b.html}, abstract = {Graph Neural Networks (GNNs) have achieved remarkable performance in modeling graphs for various applications. However, most existing GNNs assume the graphs exhibit strong homophily in node labels, i.e., nodes with similar labels are connected in the graphs. They fail to generalize to heterophilic graphs where linked nodes may have dissimilar labels and attributes. Therefore, in this paper, we investigate a novel framework that performs well on graphs with either homophily or heterophily. More specifically, we propose a label-wise message passing mechanism to avoid the negative effects caused by aggregating dissimilar node representations and preserve the heterophilic contexts for representation learning. We further propose a bi-level optimization method to automatically select the model for graphs with homophily/heterophily. Theoretical analysis and extensive experiments demonstrate the effectiveness of our proposed framework for node classification on both homophilic and heterophilic graphs.} }
Endnote
%0 Conference Paper %T Label-Wise Graph Convolutional Network for Heterophilic Graphs %A Enyan Dai %A Shijie Zhou %A Zhimeng Guo %A Suhang Wang %B Proceedings of the First Learning on Graphs Conference %C Proceedings of Machine Learning Research %D 2022 %E Bastian Rieck %E Razvan Pascanu %F pmlr-v198-dai22b %I PMLR %P 26:1--26:21 %U https://proceedings.mlr.press/v198/dai22b.html %V 198 %X Graph Neural Networks (GNNs) have achieved remarkable performance in modeling graphs for various applications. However, most existing GNNs assume the graphs exhibit strong homophily in node labels, i.e., nodes with similar labels are connected in the graphs. They fail to generalize to heterophilic graphs where linked nodes may have dissimilar labels and attributes. Therefore, in this paper, we investigate a novel framework that performs well on graphs with either homophily or heterophily. More specifically, we propose a label-wise message passing mechanism to avoid the negative effects caused by aggregating dissimilar node representations and preserve the heterophilic contexts for representation learning. We further propose a bi-level optimization method to automatically select the model for graphs with homophily/heterophily. Theoretical analysis and extensive experiments demonstrate the effectiveness of our proposed framework for node classification on both homophilic and heterophilic graphs.
APA
Dai, E., Zhou, S., Guo, Z. & Wang, S.. (2022). Label-Wise Graph Convolutional Network for Heterophilic Graphs. Proceedings of the First Learning on Graphs Conference, in Proceedings of Machine Learning Research 198:26:1-26:21 Available from https://proceedings.mlr.press/v198/dai22b.html.

Related Material