基于超邻接图的异质信息网络表征学习

doi:10.19678/j.issn.1000-3428.0065807

摘要/Abstract

摘要：

异质信息网络往往包含不同类型的节点和关系，丰富的语义信息和复杂的关系对目前异质信息网络的表征学习提出了巨大的挑战。现有多数方法通常使用预定义的元路径来捕获异质的语义信息和结构信息，但成本高、覆盖率低，且不能准确有效地捕获和学习有影响力的高阶邻居节点。提出HIN-HG模型来解决以上问题。HIN-HG通过生成异质信息网络的超邻接图来准确有效地捕获对目标节点有影响力的邻居节点，并使用带有多通道机制的卷积神经网络聚合在不同关系下的不同类型的邻居节点。HIN-HG可以自动地学习不同邻居节点和元路径的权重而无须进行手动指定，同时可以捕获全图范围内和目标节点相似的节点作为高阶邻居，并通过信息传播有效地更新目标节点的表征。在DBLP、ACM和IMDB真实数据集上的实验结果表明，在节点分类任务中，HIN-HG较HAN、GTN、HGSL等前沿的异质信息网络表征学习方法性能更优，Macro-F1和Micro-F1多分类评估指标平均提高5.6和5.7个百分点，提高了节点分类的准确性和有效性。

关键词: 异质信息网络, 元路径, 邻域聚合, 表征学习, 图卷积

Abstract:

Heterogeneous Information Network(HIN) typically contains different types of nodes and interactions. Richer semantic information and complex relationships have posed significant challenges to current representation learning in HINs. Although most existing approaches typically use predefined meta-paths to capture heterogeneous semantic and structural information, they suffer from high cost and low coverage. In addition, most existing methods cannot precisely and effectively capture and learn influential high-order neighbor nodes. Accordingly, this study attempts to address the problems of meta-paths and influential high-order neighbor nodes with a proposed original HIN-HG model. HIN-HG generates a hyperadjacency graph of the HIN, precisely and effectively capturing the influential neighbor nodes of the target nodes. Then, convolutional neural networks are adopted with a multichannel mechanism to aggregate different types of neighbor nodes under different relationships. HIN-HG can automatically learn the weights of different neighbor nodes and meta-paths without manually specifying them. Meanwhile, nodes similar to the target node can be captured in the entire graph as higher-order neighbor nodes and the representation of the target node can be effectively updated through information propagation. The experimental results of HIN-HG on three real datasets-DBLP, ACM, and IMDB demonstrate the improved performance of HIN-HG compared with state-of-the-art methods in HIN representation learning, including HAN, GTN, and HGSL. HIN-HG exhibits improved accuracy of node classification by 5.6 and 5.7 percentage points on average in the multiple classification evaluation indices Macro-F1 and Micro-F1, respectively, thus improving the accuracy and effectiveness of node classification.

Key words: Heterogeneous Information Network(HIN), meta-path, neighborhood aggregation, representation learning, graph convolution

杨彬, 王轶彤. 基于超邻接图的异质信息网络表征学习[J]. 计算机工程, 2023, 49(10): 13-21.

Bin YANG, Yitong WANG. Representation Learning in Heterogeneous Information Network Based on Hyper Adjacency Graph[J]. Computer Engineering, 2023, 49(10): 13-21.

https://www.ecice06.com/CN/Y2023/V49/I10/13

图/表 8

图1 本文方法总体框架

Fig.1 The overall framework of the proposed method

图2 节点图中不同类型边的权重

Fig.2 The weights of different types of edges in node graph

图3 语义图中不同元路径的权重

Fig.3 The weights of different meta-paths in semantic graph

图4 参数敏感性实验结果

Fig.4 Results of parameter sensitivity experiment

参考文献 35

1	SHI C , LI Y T , ZHANG J W , et al. A survey of heterogeneous information network analysis. IEEE Transactions on Knowledge and Data Engineering, 2017, 29 (1): 17- 37. doi: 10.1109/TKDE.2016.2598561
2	YANG B, WANG Y T. Representation learning in heterogeneous information networks based on hyper adjacency matrix[C]//Proceedings of International Conference on Database Systems for Advanced Applications. Berlin, Germany: Springer, 2022: 747-755.
3	CHEN J, MA T F, XIAO C. FastGCN: fast learning with graph convolutional networks via importance sampling[EB/OL]. [2022-10-12]. https://arxiv.org/abs/1801.10247.
4	DONG Y X, TANG J, WU S, et al. Link prediction and recommendation across heterogeneous social networks[C]//Proceedings of the 12th International Conference on Data Mining. Washington D.C., USA: IEEE Press, 2013: 181-190.
5	尹赢, 吉立新, 程晓涛, 等. 基于同质子图变换的异质网络表示学习. 计算机工程, 2019, 45 (11): 204- 212. doi: 10.3778/j.issn.1002-8331.1812-0254
	YIN Y , JI L X , CHENG X T , et al. Heterogeneous network representation learning based on homogeneous subgraghs transformation. Computer Engineering, 2019, 45 (11): 204- 212. doi: 10.3778/j.issn.1002-8331.1812-0254
6	郭振宏, 李海峰. 异质信息网络中演员合作关系的链路预测. 计算机工程, 2017, 43 (1): 219- 225. doi: 10.3969/j.issn.1000-3428.2017.01.038
	GUO Z H , LI H F . Link prediction of actor cooperation relationship in heterogeneous information network. Computer Engineering, 2017, 43 (1): 219- 225. doi: 10.3969/j.issn.1000-3428.2017.01.038
7	VAN DEN BERG R, KIPF T N, WELLING M. Graph convolutional matrix completion[EB/OL]. [2022-10-12]. https://arxiv.org/abs/1706.02263.
8	MONTI F, BRONSTEIN M M, BRESSON X. Geometric matrix completion with recurrent multi-graph neural networks[EB/OL]. [2022-10-12]. https://arxiv.org/abs/1704.06803.
9	SALAMAT A , LUO X , JAFARI A . HeteroGraphRec: a heterogeneous graph-based neural networks for social recommendations. Knowledge-Based Systems, 2021, 217, 106817. doi: 10.1016/j.knosys.2021.106817
10	BHAGAT S, CORMODE G, MUTHUKRISHNAN S. Node classification in social networks[M]//AGGARWAL C. Social network data analytics. Berlin, Germany: Springer, 2011: 115-148.
11	朱旭东, 熊贇. 基于多层次注意力与图模型的图像多标签分类算法. 计算机工程, 2022, 48 (4): 173-178, 190. URL
	ZHU X D , XIONG Y . Multi-label image classification algorithm based on multi-scale attention and graph model. Computer Engineering, 2022, 48 (4): 173-178, 190. URL
12	范智华, 黄大科, 李涓子, 等. 基于时间序列与多部图的演化聚类动态分析. 计算机工程, 2005, 31 (14): 156- 158.
	FAN Z H , HUANG D K , LI J Z , et al. Morphing cluster dynamics analysis based on time series and multipartite graph. Computer Engineering, 2005, 31 (14): 156- 158.
13	LI X, WU Y, ESTER M, et al. Semi-supervised clustering in attributed heterogeneous information networks[C]//Proceedings of the 26th International Conference on World Wide Web. New York, USA: ACM Press, 2017: 1621-1629.
14	KIPF T N, WELLING M. Variational graph auto-encoders[EB/OL]. [2022-10-12]. https://arxiv.org/abs/1611.07308.
15	HASAN M, MOHAMMAD A. Link prediction using supervised learning[EB/OL]. [2022-10-12]. https://www.researchgate.net/publication/200110721_Link_Prediction_using_Supervised_Learning.
16	徐上上, 孙福振, 王绍卿, 等. 基于图神经网络的异构信任推荐算法. 计算机工程, 2022, 48 (9): 89-95, 104. URL
	XU S S , SUN F Z , WANG S Q , et al. Heterogeneous trust recommendation algorithm based on graph neural networks. Computer Engineering, 2022, 48 (9): 89-95, 104. URL
17	YANG C, XIAO Y X, ZHANG Y, et al. Heterogeneous network representation learning: a unified framework with survey and benchmark[EB/OL]. [2022-10-12]. https://arxiv.org/abs/2004.00216.
18	DONG Y X, CHAWLA N V, SWAMI A. metapath2vec: scalable representation learning for heterogeneous networks[C]//Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, USA: ACM Press, 2017: 135-144.
19	FU T Y, LEE W C, LEI Z. HIN2Vec: explore meta-paths in heterogeneous information networks for representation learning[C]//Proceedings of 2017 ACM Conference on Information and Knowledge Management. New York, USA: ACM Press, 2017: 1797-1806.
20	YANG C , XIAO Y X , ZHANG Y , et al. Heterogeneous network representation learning: a unified framework with survey and benchmark. IEEE Transactions on Knowledge and Data Engineering, 2022, 34 (10): 4854- 4873. doi: 10.1109/TKDE.2020.3045924
21	WANG X, JI H Y, SHI C, et al. Heterogeneous graph attention network[C]//Proceedings of the 28th International Conference on World Wide Web. New York, USA: ACM Press, 2019: 2022-2032.
22	ZHU Z H, FAN X X, CHU X K, et al. HGCN: a heterogeneous graph convolutional network-based deep learning model toward collective classification[C]//Proceedings of the 16th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, USA: ACM Press, 2020: 1161-1171.
23	PEROZZI B, AL-RFOU R, SKIENA S. DeepWalk: online learning of social representations[C]//Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, USA: ACM Press, 2014: 701-710.
24	RONG X. word2vec parameter learning explained[EB/OL]. [2022-10-12]. https://arxiv.org/abs/1411.2738.
25	GROVER A, LESKOVEC J. node2vec: scalable feature learning for networks[C]//Proceedings of International Conference on Knowledge Discovery and Data Mining, New York, USA: ACM Press, 2016: 855-864.
26	TANG J, QU M, WANG M Z, et al. LINE: large-scale information network embedding[C]//Proceedings of the 24th International Conference on World Wide Web. New York, USA: ACM Press, 2015: 1067-1077.
27	WANG D X, CUI P, ZHU W W. Structural deep network embedding[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, USA: ACM Press, 2016: 1225-1234.
28	KIPF T N, WELLING M. Semi-supervised classification with graph convolutional networks[EB/OL]. [2022-10-12]. https://arxiv.org/abs/1609.02907.
29	VELIČKOVIĆ P, CUCURULL G, CASANOVA A, et al. Graph attention networks[EB/OL]. [2022-10-12]. https://arxiv.org/abs/1710.10903.
30	HAMILTON W L, YING R, LESKOVEC J. Inductive representation learning on large graphs[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems. New York, USA: ACM Press, 2017: 1025-1035.
31	SCHLICHTKRULL M, KIPF T N, BLOEM P, et al. Modeling relational data with graph convolutional networks[C]//Proceedings of European Semantic Web Conference. Berlin, Germany: Springer, 2018: 593-607.
32	ZHANG C X, SONG D J, HUANG C, et al. Heterogeneous graph neural network[C]//Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, USA: ACM Press, 2019: 793-803.
33	LU Y F, FANG Y, SHI C. Meta-learning on heterogeneous information networks for cold-start recommendation[C]//Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, USA: ACM Press, 2020: 1563-1573.
34	YUN S, JEONG M, KIM R, et al. Graph Transformer networks[EB/OL]. [2022-10-12]. https://arxiv.org/abs/1911.06455.
35	ZHAO J N , WANG X A , SHI C A , et al. Heterogeneous graph structure learning for graph neural networks. Proceedings of AAAI Conference on Artificial Intelligence, 2021, 35 (5): 4697- 4705. doi: 10.1609/aaai.v35i5.16600

[1]	张洋, 刘畅, 李少青. 基于可控制性度量的图神经网络门级硬件木马检测方法[J]. 计算机工程, 2024, 50(7): 164-173.
[2]	代巍, 王丰羽, 冀常鹏. 基于情感增强与双图卷积网络的方面级情感分析[J]. 计算机工程, 2024, 50(5): 120-127.
[3]	吴正洋, 张广涛, 黄立, 汤庸. 基于异质图嵌入和会话交互的课程推荐模型[J]. 计算机工程, 2024, 50(4): 95-103.
[4]	杨春霞, 吴亚雷, 闫晗, 黄昱锟. 融合双图卷积与门控线性单元的方面级情感分析模型[J]. 计算机工程, 2024, 50(4): 141-149.
[5]	莫少聪, 陈庆锋, 谢泽, 刘春雨, 邱俊铼. 基于动态图注意力与标签传播的实体对齐[J]. 计算机工程, 2024, 50(4): 150-159.
[6]	李政学, 李枝名, 彭德中, 陈杰. 基于特征对比学习和图卷积的社交网络用户分类[J]. 计算机工程, 2024, 50(4): 258-266.
[7]	夏卫欢, 廖列法, 张守信, 张燕琴. 基于方面-词性感知的方面级情感分析[J]. 计算机工程, 2024, 50(3): 68-77.
[8]	袁文涛, 卫文韬, 高德民. 融合注意力机制的多视图卷积手势识别研究[J]. 计算机工程, 2024, 50(3): 208-215.
[9]	冉丈杰, 孙林夫, 邹益胜, 马玉麟. 基于关系学习网络的小样本知识图谱补全模型[J]. 计算机工程, 2023, 49(9): 52-59.
[10]	隋国华, 李陶然, 刘昊, 陈林, 汪卫. 基于图表示学习的领域知识图谱推理技术研究[J]. 计算机工程, 2023, 49(9): 89-98.
[11]	马建红, 龚天, 姚爽. 基于证据句与图卷积网络的文档级关系抽取[J]. 计算机工程, 2023, 49(8): 104-110.
[12]	王款, 宣士斌, 何雪东, 李紫薇, 李嘉祥. 基于交叉注意力Transformer的人体姿态估计方法[J]. 计算机工程, 2023, 49(7): 223-231.
[13]	代祖华, 刘园园, 狄世龙. 语义增强的图神经网络方面级文本情感分析[J]. 计算机工程, 2023, 49(6): 71-80.
[14]	陈昱瑾, 王晶, 武志昊, 赵耀帅, 林友芳. 基于图卷积网络融合群组关系的社会化推荐方法[J]. 计算机工程, 2023, 49(5): 112-121.
[15]	陈文轩, 曾碧, 郭植星. 融合多特征与语义图卷积网络的摔倒检测方法[J]. 计算机工程, 2023, 49(5): 277-285,294.

选择文件类型/文献管理软件名称

选择包含的内容