[1]ZHENG Shuai,PENG Zouzhang,ZHU Zhenfeng,et al.Graph disentanglement representation learning based on propagation in multiple hyperbolic spaces[J].CAAI Transactions on Intelligent Systems,2025,20(5):1243-1255.[doi:10.11992/tis.202409034]
Copy

Graph disentanglement representation learning based on propagation in multiple hyperbolic spaces

CAAI Transactions on Intelligent Systems[ISSN 1673-4785/CN 23-1538/TP] Volume: 20 Number of periods: 2025 5 Page number: 1243-1255 Column: 吴文俊人工智能科学技术奖论坛 Public date: 2025-09-05
Title:
Graph disentanglement representation learning based on propagation in multiple hyperbolic spaces
Author(s):
ZHENG Shuai1; 2;  PENG Zouzhang1; 2;  ZHU Zhenfeng1; 2;  ZHAO Yao1; 2
1. School of Computer Science and Technology, Beijing Jiaotong University, Beijing 100044, China;
2. Beijing Key Laboratory of Advanced Information Science and Network Technology, Beijing 100044, China
Keywords:
graph representation learning; graph disentanglement; hyperbolic space; graph neural networks; label propagation; mixture of experts; topology refinement; multiresolution
CLC:
TP18
DOI:
10.11992/tis.202409034
Abstract:
There are two salient issues of existing graph representation learning methods. First, there is a dearth of fine-grained neighborhood modeling, which neglects the multifaceted semantic entanglements in the neighborhood structures. Second, the spatial metric employed in graph representation learning presents a significant challenge, since Euclidean space may not constitute the optimal framework for quantifying node representations. To solve these challenges, this study proposes a novel representation propagation and prediction mechanism within multiple hyperbolic spaces, thereby achieving disentangled graph representation learning under multifaceted hyperbolic spatial metrics. Within the proposed framework, the original topological structure is iteratively refined through node representations, yielding propagation matrices embedded in a hyperbolic space. Furthermore, based on a mixture-of-experts design, hyperbolic label propagation networks at different resolutions are treated as expert networks, enabling the discovery of node connection patterns induced by different latent factors. Experimental results on multiple real-world datasets show that the proposed method achieves classification accuracies of 32.3% and 59.5% on the Squirrel and Crocodile datasets, respectively. Additionally, visualization experiments further demonstrate the effectiveness of the proposed approach.
References:
[1] 侯磊, 刘金环, 于旭, 等. 图神经网络研究综述[J]. 计算机科学, 2024, 51(6): 282-298.
HOU Lei, LIU Jinhuan, YU Xu, et al. Review of graph neural networks[J]. Computer science, 2024, 51(6): 282-298.
[2] 白雪. 基于图神经网络的推荐方法研究[D]. 太原: 山西大学, 2023: 1-46.
BAI Xue. Research on recommendation method based on graph neural network[D]. Taiyuan: Shanxi University, 2023: 1-46.
[3] 张舒菡. 基于图表示学习的中医处方推荐方法研究[D]. 北京: 北京交通大学, 2023: 1-64.
ZHANG Shuhan. Study on TCM prescription recommendation method based on graph representation learning[D]. Beijing: Beijing Jiaotong University, 2023: 1-64.
[4] 张嘉伟. 基于图结构学习的图神经网络的药物靶标预测应用[D]. 武汉: 华中农业大学, 2023: 1-58.
ZHANG Jiawei. Application of graph neural network based on graph structure learning to drug target prediction[D]. Wuhan: Huazhong Agricultural University, 2023: 1-58.
[5] 张宇轩. 基于异构图表示学习的药物—靶蛋白关联关系预测研究[D]. 长春: 东北师范大学, 2023: 1-51.
ZHANG Yuxuan. Drug-target interaction prediction via heterogeneous graph representation learning[D]. Changchun: Northeast Normal University, 2023: 1-51.
[6] 邹然, 柳杨, 李聪, 等. 图表示学习综述[J]. 北京师范大学学报(自然科学版), 2023, 59(5): 716-724.
ZOU Ran, LIU Yang, LI Cong, et al. Graph representation learning: a review[J]. Journal of Beijing Normal University (natural science), 2023, 59(5): 716-724.
[7] GILMER J, SCHOENHOLZ S S, RILEY P F, et al. Neural message passing for quantum chemistry[EB/OL]. (2017-06-12)[2024-09-26]. https://arxiv.org/abs/1704.01212v2.
[8] LIN Guangfeng, KANG Xiaobing, LIAO Kaiyang, et al. Deep graph learning for semi-supervised classification[J]. Pattern recognition, 2021, 118: 108039.
[9] WU F, SOUZA A, ZHANG Tianyi, et al. Simplifying graph convolutional networks[C]//International Conference on Machine Learning. Los Angeles: PMLR, 2019: 6861–6871.
[10] MA Jianxin, CUI Peng, KUANG Kun, et al. Disentangled graph convolutional networks[C]//International Conference on Machine Learning. Los Angeles: PMLR, 2019: 4212-4221.
[11] YANG Yiding, FENG Zunlei, SONG Mingli, et al. Factorizable graph convolutional networks[J]. Advances in neural information processing systems, 2020, 33: 20286-20296.
[12] CHAMI I, YING R, Ré C, et al. Hyperbolic graph convolutional neural networks[J]. Advances in neural information processing systems, 2019, 32: 4869-4880.
[13] LIU Qi, NICKEL M, KIELA D. Hyperbolic graph neural networks[J]. Advances in neural information processing systems, 2019, 32: 8230-8241.
[14] ZHANG Yiding, WANG Xiao, SHI Chuan, et al. Hyperbolic graph attention network[J]. IEEE transactions on big data, 2022, 8(6): 1690-1701.
[15] 刘笑梅. 双曲空间中的神经网络图表示学习[D]. 桂林: 广西师范大学, 2019: 1-46.
LIU Xiaomei. Graph representation learning with neural network in hyperbolic space[D]. Guilin: Guangxi Normal University, 2019: 1-46.
[16] 周志霞. 基于自编码器与双曲几何的网络表示学习[D]. 石家庄: 河北师范大学, 2023: 1-53.
ZHOU Zhixia. Network representation learning based on auto-encoder and hyperbolic geometry[D]. Shijiazhuang: Hebei Normal University, 2023: 1-53.
[17] XU Keyulu, HU Weihua, LESKOVEC J, et al. How powerful are graph neural networks?[C]//International Conference on Learning Representations. New Orleans: OpenReview.net, 2019: 1-17.
[18] PAGE L, BRIN S, MOTWANI R, et al. The pagerank citation ranking: bring order to the web [C]//International Conference on World Wide Web. Brisbane: Elsevier, 1998: 1-17.
[19] SPINELLI I, SCARDAPANE S, UNCINI A. Adaptive propagation graph convolutional network[J]. IEEE transactions on neural networks and learning systems, 2020, 32(10): 4755-4760.
[20] GASTEIGER J, BOJCHEVSKI A, GüNNEMANN S. Predict then propagate: graph neural networks meet personalized pagerank[C]//International Conference on Learning Representations. New Orleans: OpenReview. net, 2019: 1-15.
[21] KIPF T, WELLING M. Semi-supervised classification with graph convolutional networks[EB/OL]. (2017-02-22)[2024-05-09]. https://arxiv.org/abs/1609.02907.
[22] LI Guohao, MüLLER M, GHANEM B, et al. Training graph neural networks with 1000 layers[C]//International Conference on Machine Learning. [S. l. ]: PMLR, 2021: 6437–6449.
[23] SARA S, NICHOLAS F, GEOFFREY H. Dynamic routing between capsules[J]. Advances in neural information processing systems, 2017, 30: 3859-3869.
[24] ZHAO Tianxiang, ZHANG Xiang, WANG Suhang. Exploring edge disentanglement for node classification[C]//Proceedings of the ACM Web Conference 2022. Lyon: ACM, 2022: 1028-1036.
[25] LIU Yanbei, WANG Xiao, WU Shu, et al. Independence promoted graph disentangled networks[C]//Proceedings of the AAAI Conference on Artificial Intelligence. New York: AAAI, 2020: 4916-4923.
[26] ZHENG Shuai, ZHU Zhenfeng, LIU Zhizhe, et al. Adversarial graph disentanglement with component-specific aggregation[J]. IEEE transactions on artificial intelligence, 2024, 5(5): 2204-2216.
[27] WANG Yifan, TANG Suyao, LEI Yuntong, et al. DisenHAN: disentangled heterogeneous graph attention network for recommendation[C]//Proceedings of the 29th ACM International Conference on Information & Knowledge Management. [S. l. ]: ACM, 2020: 1605-1614.
[28] XIAO Teng, CHEN Zhengyu, GUO Zhimeng, et al. Decoupled self-supervised learning for graphs[J]. Advances in neural information processing systems, 2022, 35: 620-634.
[29] VELI?KOVI? P, CUCURULL G, CASANOVA A, et al. Graph attention networks[EB/OL]. (2018-02-04) [2024-05-09]. https://arxiv.org/abs/1710.10903.
[30] SUN Jianing, CHENG Zhaoyue, ZUBERI S, et al. HGCF: hyperbolic graph convolution networks for collaborative filtering[C]//Proceedings of the Web Conference 2021. Ljubljana: ACM, 2021: 593-601.
[31] 张依丁. 基于双曲空间的图表示学习算法研究[D]. 北京: 北京邮电大学, 2022: 1-98.
ZHANG Yiding. Research on hyperbolic graph representation learning algorithm graph representation learning algorithm based on hyperbolic space[D]. Beijing: Beijing University of Posts and Telecommunications, 2022: 1-98.
[32] GREGOR B, GARY B, OCTAVIAN G. Constant curvature graph convolutional networks[C]//International Conference on Machine Learning. [S. l. ]: PMLR, 2020: 486-496.
[33] ZHU Shichao, PAN Shirui, ZHOU Chuan, et al. Graph geometry interaction learning[J]. Advances in neural information processing systems, 2020, 33: 7548-7558.
[34] ZHOU Zhiheng, WANG Tao, HOU Linfang, et al. Decoupled hyperbolic graph attention network for modeling substitutable and complementary item relationships[C]//Proceedings of the 31st ACM International Conference on Information & Knowledge Management. Atlanta: ACM, 2022: 2763–2772.
[35] LIU Meng, GAO Hongyang, JI Shuiwang. Towards deeper graph neural networks[C]//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. [S. l. ]: ACM, 2020: 338-348.
[36] QIAN Huang, HORACE H, ABHAY S, et al. Combining label propagation and simple models out-performs graph neural networks[C]//International Conference on Learning Representations. [S. l. ]: OpenReview.net, 2021: 1-21.
[37] DONG Hande, CHEN Jiawei, FENG Fuli, et al. On the equivalence of decoupled graph convolution network and label propagation[C]//Proceedings of the Web Conference 2021. Ljubljana: ACM, 2021: 3651-3662.
[38] DING Kaize, WANG Jianling, CAVERLEE J, et al. Meta propagation networks for graph few-shot semi-supervised learning[C]//AAAI Conference on Artificial Intelligence, 2022 , 36(6): 6524-6531.
[39] 王振胜. 基于特征平滑的图神经网络在节点分类中的应用研究[D]. 北京: 北京化工大学, 2023: 9-41.
WANG Zhensheng. Research on the application of graph neural networks with feature smoothing in node classification[D]. Beijing: Beijing University of Chemical Technology, 2023: 9-41.
[40] SHEN Qinghua, REN Weijieying, QIN Wei. Graph relation aware continual learning [EB/OL]. (2023-08-16) [2024-05-09]. https://arxiv.org/abs/2308.08259.
[41] HAMILTON W L, YING R, LESKOVEC J. Inductive representation learning on large graphs[J]. Advances in neural information processing systems, 2017, 30: 1-11.
[42] YANG Ling, TIAN Ye, XU Minkai, et al. VQGraph: rethinking graph representation space for bridging GNNs and MLPs[C]//International Conference on Learning Representations. Vienna: OpenReview.net, 2024: 1-20.
[43] XIAO Teng, ZHU Huaisheng, CHEN Zhengyu, et al. Simple and asymmetric graph contrastive learning without augmentations[J]. Advances in neural information processing systems, 2024, 36: 1-24.
[44] CHIEN E, PENG Jianhao, LI Pan, et al. Adaptive universal generalized PageRank graph neural network[C]//International Conference on Learning Representations. [S. l. ]: OpenReview.net, 2021: 1-24.
[45] YU Tao, DE SA C. Random Laplacian features for learning with hyperbolic space[C]//International Conference on Learning Representations. Kigali: OpenReview.net, 2023: 1-23.
[46] ZHU Jiong, YAN Yujun, ZHAO Lingxiao, et al. Beyond homophily in graph neural networks: current limitations and effective designs[J]. Advances in neural information processing systems, 2020, 33: 7793-7804.
[47] WANG Xiyuan, ZHANG Muhan. How powerful are spectral graph neural networks[C]//International Conference on Machine Learning. Baltimore: PMLR, 2022: 23341-23362.
[48] PEI Hongbin, WEI Bingzhen, CHANG K, et al. Geom-GCN: geometric graph convolutional networks [C]//International Conference on Learning Representations. [S. l. ]: OpenReview.net, 2020: 1-12.
Similar References:

Memo

-

Last Update: 2025-09-05

Copyright © CAAI Transactions on Intelligent Systems