Multi-level disentanglement graph neural network

被引:5
|
作者
Wu, Lirong [1 ,2 ]
Lin, Haitao [2 ]
Xia, Jun [2 ]
Tan, Cheng [2 ]
Li, Stan Z. [2 ]
机构
[1] Zhejiang Univ, Coll Comp Sci & Technol, Hangzhou 310058, Zhejiang, Peoples R China
[2] Westlake Univ, Sch Engn, AI Lab, Hangzhou 310024, Zhejiang, Peoples R China
来源
NEURAL COMPUTING & APPLICATIONS | 2022年 / 34卷 / 11期
基金
中国国家自然科学基金;
关键词
Graph neural networks; Disentanglement; Relation learning; Semi-supervised learning;
D O I
10.1007/s00521-022-06930-1
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Real-world graphs are generally generated from highly entangled latent factors. However, existing deep learning methods for graph-structured data often ignore such entanglement and simply denote the heterogeneous relations between entities as binary edges. In this paper, we propose a novel Multi-level Disentanglement Graph Neural Network (MD-GNN), a unified framework that simultaneously implements edge-level, attribute-level, and node-level disentanglement in an end-to-end manner. MD-GNN takes the original graph structure and node attributes as input and outputs multiple disentangled relation graphs and disentangled node representations. Specifically, MD-GNN first disentangles the original graph structure into multiple relation graphs, each of which corresponds to a latent and disentangled relation among entities. The input node attributes are then propagated in the corresponding relation graph through a multi-hop diffusion mechanism to capture long-range dependencies between entities, and finally the disentangled node representations are obtained through information aggregation and merging. Extensive experiments on synthetic and real-world datasets have shown qualitatively and quantitatively that MD-GNN yields truly encouraging results in terms of disentanglement and also serves well as a general GNN framework for downstream tasks. Code has been made available at: https://github.com/LirongWu/MD-GNN.
引用
收藏
页码:9087 / 9101
页数:15
相关论文
共 50 条
  • [1] Multi-level disentanglement graph neural network
    Lirong Wu
    Haitao Lin
    Jun Xia
    Cheng Tan
    Stan Z. Li
    Neural Computing and Applications, 2022, 34 : 9087 - 9101
  • [2] Multi-level graph neural network for text sentiment analysis
    Liao, Wenxiong
    Zeng, Bi
    Liu, Jianqi
    Wei, Pengfei
    Cheng, Xiaochun
    Zhang, Weiwen
    COMPUTERS & ELECTRICAL ENGINEERING, 2021, 92
  • [3] A spatiotemporal motion prediction network based on multi-level feature disentanglement
    Chen, Suting
    Bo, Yewen
    Wu, Xu
    IMAGE AND VISION COMPUTING, 2024, 146
  • [4] Multi-level graph convolutional recurrent neural network for semantic image segmentation
    Jiang, Dingchao
    Qu, Hua
    Zhao, Jihong
    Zhao, Jianlong
    Liang, Wei
    TELECOMMUNICATION SYSTEMS, 2021, 77 (03) : 563 - 576
  • [5] Multi-level graph convolutional recurrent neural network for semantic image segmentation
    Dingchao Jiang
    Hua Qu
    Jihong Zhao
    Jianlong Zhao
    Wei Liang
    Telecommunication Systems, 2021, 77 : 563 - 576
  • [6] Inferring Users' Social Roles with a Multi-Level Graph Neural Network Model
    Zhang, Chunrui
    Wang, Shen
    Zhan, Dechen
    Yin, Mingyong
    Lou, Fang
    ENTROPY, 2021, 23 (11)
  • [7] A multi-level probabilistic neural network
    Zong, Ning
    Hong, Xia
    ADVANCES IN NEURAL NETWORKS - ISNN 2007, PT 2, PROCEEDINGS, 2007, 4492 : 516 - +
  • [8] Multi-Level Graph Neural Network With Sparsity Pooling for Recognizing Parkinson’s Disease
    Zhang, Xiaobo
    Zhou, Yuxin
    Lu, Zhijie
    Zhai, Donghai
    Luo, Haonan
    Li, Tianrui
    Li, Yang
    IEEE TRANSACTIONS ON NEURAL SYSTEMS AND REHABILITATION ENGINEERING, 2023, 31 : 4459 - 4469
  • [9] Topic-Enhanced Multi-level Graph Neural Network for Session-Based Recommendation
    Tang G.
    Zhu X.
    Moshi Shibie yu Rengong Zhineng/Pattern Recognition and Artificial Intelligence, 2023, 36 (02): : 174 - 186
  • [10] Multi-level category-aware graph neural network for session-based recommendation
    Zhang, Zhu
    Yang, Bo
    Xu, Hao
    Hu, Wang
    EXPERT SYSTEMS WITH APPLICATIONS, 2024, 242