本项目用于整理异质图神经网络(HGNN)相关数据、算法、其他资料。
非使用GNN方法的异质图相关工作(尤其是推荐系统和知识图谱相关工作)暂不列入。暂时主要考虑创新点在GNN方面的(如NLP领域主要创新点在如何建图方面的,暂不列入)。
作者主要用PyTorch和PyG。
超过5M的文件都存储在了百度网盘上,以方便大陆用户下载:
链接:https://pan.baidu.com/s/1P1X9LjT85DU9OBPAr6l3bg
提取码:ea7u
简单介绍:
数据集名称 | 下载和预处理方式 | 出处 | 常用任务 |
---|---|---|---|
ogbn-mag | PyG OGM_MAG ogb |
MAG数据集的子集 | 节点分类 |
MAG | 已停用 | Microsoft academic graph: When experts are not enough. | |
AMiner (metapath2vec) | PyG(百度网盘) | metapath2vec: Scalable Representation Learning for Heterogeneous Networks | 节点分类 |
AMinerNetwork | https://www.aminer.org/aminernetwork | A multilayered approach for link prediction in heterogeneous complex networks | 链路预测 |
DBIS | https://ericdongyx.github.io/metapath2vec/m2v.html (百度网盘和HGNN_Collection/load_data/dbis_pyg.py) | metapath2vec: Scalable Representation Learning for Heterogeneous Networks | 节点相似度 |
DBLP (MAGNN) | PyG (DBLP和HGBDataset)(百度网盘) | MAGNN: Metapath Aggregated Graph Neural Network for Heterogeneous Graph Embedding | 节点分类 |
IMDB (MAGNN) | PyG(百度网盘) | MAGNN: Metapath Aggregated Graph Neural Network for Heterogeneous Graph Embedding | 节点分类 |
IMDB (Simple-HGN) | PyG HGBDataset | Are We Really Making Much Progress? Revisiting, Benchmarking, and Refining Heterogeneous Graph Neural Networks | 节点分类 |
LastFM | PyG(百度网盘) | MAGNN: Metapath Aggregated Graph Neural Network for Heterogeneous Graph Embedding | 链路预测 |
MovieLens (PyG) | PyG | https://movielens.org/ | |
ACM | PyG HGBDataset | Are We Really Making Much Progress? Revisiting, Benchmarking, and Refining Heterogeneous Graph Neural Networks | 节点分类 |
Freebase | PyG HGBDataset | Are We Really Making Much Progress? Revisiting, Benchmarking, and Refining Heterogeneous Graph Neural Networks | 节点分类 |
ogbl-biokg | https://ogb.stanford.edu/docs/linkprop/ | 链路预测 |
详细介绍:
《HGNN图数据集》
相关介绍博文: 异质图神经网络(HGNN)常用数据集信息统计(持续更新ing...)_诸神缄默不语的博客-CSDN博客_异质图数据集 PyG (PyTorch Geometric) Dropbox系图数据集无法下载的解决方案(AMiner, DBLP, IMDB, LastFM)(持续更新ing...)
2024年
KG嵌入:
2023年
通用节点嵌入:
- (WWW) A Post-Training Framework for Improving Heterogeneous Graph Neural Networks
- (Journal of Computational Science) Multi-view contrastive learning for multilayer network embedding
- (Entropy) Unsupervised Embedding Learning for Large-Scale Heterogeneous Networks Based on Metapath Graph Sampling
- (Scientific Reports) A multi-view contrastive learning for heterogeneous network embedding
- HINormer: Representation Learning On Heterogeneous Information Networks with Graph Transformer
- Heterophily-Aware Graph Attention Network
动态图节点嵌入:
- (Neurocomputing) Dynamic heterogeneous graph representation learning with neighborhood type modeling
graph rewiring:
- (WWW) Homophily-oriented Heterogeneous Graph Rewiring:同配性+异质性
graphlet和orbit计数:
交叉学科:
- (Phys. Rev. E) Deterministic, quenched and annealed parameter estimation for heterogeneous network models:计量经济学+物理
2022年
综述:
- (Artificial Intelligence Review) Heterogeneous graph neural networks analysis: a survey of techniques, evaluations and applications
通用节点嵌入:
- (World Wide Web) HGNN-ETA: Heterogeneous graph neural network enriched with text attribute:利用节点的文本特征信息
- (World Wide Web) How the four-nodes motifs work in heterogeneous node representation?
- (CIKM) SplitGNN: Splitting GNN for Node Classification with Heterogeneous Attention:联邦学习
- (SDM) Structure-Enhanced Heterogeneous Graph Contrastive Learning:提出STENCIL模型,跨视图(保证视图之间一致性最大化)(基于metapaths构建视图(metapath实例起终点构成的同质图),最大化同一节点在不同视图上嵌入的相似性,将各视图的嵌入attentively聚合)+对比学习+结构嵌入
- (Transactions on Big Data) A Survey on Heterogeneous Graph Embedding: Methods, Techniques, Applications and Sources:综述
- (Knowledge-Based Systems) Megnn: Meta-path extracted graph neural network for heterogeneous graph representation learning:自动提取metapaths+可解释性
- (IEEE Transactions on Knowledge and Data Engineering) Heterogeneous Graph Representation Learning with Relation Awareness:提出R-HGNN模型
- (IEEE Transactions on Knowledge and Data Engineering) mg2vec: Learning Relationship-Preserving Heterogeneous Graph Representations via Metagraph Embedding:meta graph
- (IEEE Transactions on Knowledge and Data Engineering) RHINE: Relation Structure-Aware Heterogeneous Information Network Embedding
- (IEEE Transactions on Knowledge and Data Engineering) Explicit Message-Passing Heterogeneous Graph Neural Network:提出EMP模型。本文认为传统用metapaths将异质图构建为同质图的方法是隐式的,而EMP模型显式信息传递
- (南洋理工大学硕士论文) Contrastive learning for heterogeneous graph neural networks
- (Neural Networks) Latent neighborhood-based heterogeneous graph representation:随机游走增强邻居+GTN学习metapaths
- (IEEE Transactions on Neural Networks and Learning Systems) Learning Knowledge Graph Embedding With Heterogeneous Relation Attention Networks
- (Appl. Sci.) MBHAN: Motif-Based Heterogeneous Graph Attention Network:基于motif
- (Data Mining and Knowledge Discovery) Personalised meta-path generation for heterogeneous graph neural networks:提出PM-HGNN模型,强化学习(将找metapaths视作马尔科夫决策过程)
- (Machine Learning) Heterogeneous graph embedding with single-level aggregation and infomax encoding:提出无监督HGNN模型HIME,对节点特征应用MLP(每一种节点用一个模型),直接进行聚合。损失函数鼓励邻居相近+信息最大化
- (International Journal of Machine Learning and Cybernetics) Multiple heterogeneous network representation learning based on multi-granularity fusion:将结构(一阶邻居)和语义(meta-path)视作不同粒度,结合来学习表征
- (Journal of Applied Mathematics) Classification Algorithm for Heterogeneous Network Data Streams Based on Big Data Active Learning
- (IEEE Access) Siamese Network Based Multiscale Self-Supervised Heterogeneous Graph Representation Learning:提出SNMH模型,自监督学习+对比学习(metapaths和one-hop)+孪生神经网络
- (ICCWAMTIP) Relation Heterogeneous Graph Neural Network
- (ICDMW) Graph Convolutional Neural Network based on the Combination of Multiple Heterogeneous Graphs
- Simple and Efficient Heterogeneous Graph Neural Network:提出SeHGNN模型,预处理+无参+轻量级
- Relation Embedding based Graph Neural Networks for Handling Heterogeneous Graph:不用metapaths
- Descent Steps of a Relation-Aware Energy Produce Heterogeneous Graph Neural Networks:关注过平滑问题
- Heterogeneous Graph Neural Networks using Self-supervised Reciprocally Contrastive Learning:提出HGCL模型,自监督学习+对比学习(节点属性和拓扑结构)
- Heterogeneous Graph Contrastive Multi-view Learning:提出HGCML模型,关注减轻对比学习(metapaths)中的采样偏差
- Heterogeneous Graph Masked Autoencoders:提出HGMAE模型,auto encoder
- Meta-node: A Concise Approach to Effectively Learn Complex Relationships in Heterogeneous Graphs:不用预定义的meta-paths或meta-graphs
预训练:
multiplex graph:
- (KDD) Multiplex Heterogeneous Graph Convolutional Network
- (ijtr) DSMN: A New Approach for Link Prediction in Multilplex Networks
小样本学习:
- (KDD) Few-shot Heterogeneous Graph Learning via Cross-domain Knowledge Transfer
AI安全:
- (AAAI) Robust Heterogeneous Graph Neural Networks against Adversarial Attacks
知识蒸馏:
- (WWW) Collaborative Knowledge Distillation for Heterogeneous Information Network Embedding
- (Neurocomputing) HIRE: Distilling high-order relational knowledge from heterogeneous graph neural networks
匹配节点:
- (AAAI) From One to All: Learning to Match Heterogeneous and Partially Overlapped Graphs
链路预测:
2021年
通用节点嵌入:
- (KDD) Are we really making much progress? Revisiting, benchmarking, and refining heterogeneous graph neural networks:先把近年多篇论文(HAN、GTN、RSHN、HetGNN、MAGNN、HGT、HetSANN、RGCN、GATNE、KGCN、KAGT)喷了一遍,然后提出Simple-HGN模型(参考博文:Re10:读论文 Are we really making much progress? Revisiting, benchmarking, and refining heterogeneous gr_诸神缄默不语的博客-CSDN博客)
- (KDD) Self-supervised Heterogeneous Graph Neural Network with Co-contrastive Learning:提出HeCo模型,跨视图(metapath和network)+对比学习+自监督学习
- (KDD) HGK-GNN: Heterogeneous Graph Kernel based Graph Neural Networks:引入HGK
- (KDD) DiffMG: Differentiable Meta Graph Search for Heterogeneous Graph Neural Networks:NAS
- (AAAI) Heterogeneous graph structure learning for graph neural networks
- (WWW) Heterogeneous Graph Neural Network via Attribute Completion:提出HGNN-AC模型,端到端地同时学习节点特征补全和GNN学习
- (IJCAI) Heterogeneous Graph Information Bottleneck:提出HGIB模型,无监督学习+信息论
- (ECML PKDD) Multi-view Self-supervised Heterogeneous Graph Embedding:提出MVSE模型,用基于metapaths的多视图实现自监督学习
- (IEEE Transactions on Knowledge and Data Engineering) Heterogeneous Graph Propagation Network:提出HPN模型,缓解HGNN中的confusion phenomenon问题(类似同质GNN中的过平滑问题)
- (IEEE Transactions on Knowledge and Data Engineering) Interpretable and Efficient Heterogeneous Graph Convolutional Network:提出ie-HGCN模型,可以广泛评估各种metapaths
- (IEEE Transactions on Knowledge and Data Engineering) Higher-Order Attribute-Enhancing Heterogeneous Graph Neural Networks:提出HAE模型,使用了meta-graphs
- (IEEE Transactions on Knowledge and Data Engineering) HGATE: Heterogeneous Graph Attention Auto-Encoders:无监督学习+auto encoder
- (IEEE Transactions on Knowledge and Data Engineering) Walking with Attention: Self-guided Walking for Heterogeneous Graph Embedding:提出SILK模型,基于随机游走
- (Big Data Mining and Analytics) Attention-aware heterogeneous graph neural network:提出AHNN模型,metapath+attention
- (ICME) Revisiting Graph Neural Networks for Node Classification in Heterogeneous Graphs:关注过拟合问题
- (ICLR 2022被拒+撤回投稿) R-GSN: The Relation-based Graph Similar Network for Heterogeneous Graph:不用metapaths,据称在ogbn-mag图上达到了SOTA效果
动态图:
- (ECML PKDD) Dynamic Heterogeneous Graph Embedding via Heterogeneous Hawkes Process:霍克斯过程
超图:
- (WSDM) Heterogeneous Hypergraph Embedding for Graph Classification
社区检测:
- (CIKM) Detecting Communities from Heterogeneous Graphs: A Context Path-based Graph Neural Network Model:提出CP-GNN模型
图生成:
- (ICDM) Deep Generation of Heterogeneous Networks:提出HGEN模型
2020年
通用节点嵌入:
- (IJCAI) Heterogeneous Network Representation Learning:综述
- (AAAI) An Attention-based Graph Neural Network for Heterogeneous Structural Learning:提出HetSANN模型,基于关系类型,建立attention机制,实现节点信息聚合,不使用metapath(参考博文:Re22:读论文 HetSANN An Attention-based Graph Neural Network for Heterogeneous Structural Learning_诸神缄默不语的博客-CSDN博客)
- (WWW) MAGNN: Metapath Aggregated Graph Neural Network for Heterogeneous Graph Embedding:首先对节点特征进行转换,然后聚合metapath内部信息,然后聚合各metapath的信息
- (WWW) Heterogeneous Graph Transformer:提出HGT模型,把整个Transformer结构改到图上,这种感觉
- Hybrid Micro/Macro Level Convolution for Heterogeneous Graph Learning:提出HGConv模型
- Reinforcement Learning Enhanced Heterogeneous Graph Neural Network:提出HGRL模型,强化学习(将找metapaths视作马尔科夫决策过程)
- (被ICLR 2021拒了) Scalable Graph Neural Networks for Heterogeneous Graphs:提出NARS模型,关注scalability问题
动态图:
- (ECIR) Dynamic Heterogeneous Graph Embedding Using Hierarchical Attentions:提出DyHAN模型(实验仅做了链路预测)
- (ICKG) Heterogeneous Dynamic Graph Attention Network:提出HDGAN模型
- Meta Graph Attention on Heterogeneous Graph with Node-Edge Co-evolution:提出CoMGNN和ST-CoMGNN模型
multiplex graph:
- (AAAI) Unsupervised Attributed Multiplex Network Embedding:提出DMGI模型
多模态:
- (KDD) HGMF: Heterogeneous Graph-based Fusion for Multimodal Data with Incompleteness
anchor link prediction任务:
- (AAAI) Type-Aware Anchor Link Prediction across Heterogeneous Networks Based on Graph Attention Network
2019年
通用节点嵌入:
- (WWW) Heterogeneous Graph Attention Network:提出HAN模型,先基于metapath attentively聚合节点信息,然后attentively聚合metapath信息
- (NeurIPS) Graph Transformer Networks:提出GTN模型,自动学习metapaths
- (KDD) Heterogeneous Graph Neural Network:提出HetGNN模型,用RWR抽样异质邻居,按节点类型分类,然后用聚合
- (KDD) Representation Learning for Attributed Multiplex Heterogeneous Network:提出GATNE模型
- (KDD) Adversarial Learning on Heterogeneous Information Networks:提出HetGAN模型
- (ICDM) Relation Structure-Aware Heterogeneous Graph Neural Network:提出RSHN模型,用coarsened line graph先获得边特征,然后传播节点和边特征
- (CIKM) BHIN2vec: Balancing the Type of Relation in Heterogeneous Information Network:提出BHIN2vec模型,随机游走+skip gram+多任务,解决HIN中不同种类边数不平衡的问题
- Heterogeneous Deep Graph Infomax:提出HDGI模型,信息论+无监督学习
图匹配:
- (IJCAI) Heterogeneous Graph Matching Networks for Unknown Malware Detection
底层编译技术:
- (PACT) Gluon-Async: A Bulk-Asynchronous System for Distributed and Heterogeneous Graph Analytics
2018年
通用节点嵌入:
- (CIKM) Are Meta-Paths Necessary?: Revisiting Heterogeneous Graph Embeddings
- (KDD) Easing Embedding Learning by Comprehensive Transcription of Heterogeneous Information Networks
- (TKDE) Heterogeneous Information Network Embedding for Recommendation:提出HERec模型,将异质图转换为metapath-based graphs,用skip-gram嵌入
- (ESWC) Modeling Relational Data with Graph Convolutional Networks:提出RGCN模型
- (SIAM) AspEm: Embedding Learning by Aspects in Heterogeneous Information Networks
链路预测:
2017年
通用节点嵌入:
- (KDD) metapath2vec: Scalable Representation Learning for Heterogeneous Networks:用基于metapath的随机游走来构建邻居,然后用类似word2vec的逻辑来实现节点表征(类似DeepWalk)(参考博文:Re31:读论文 metapath2vec: Scalable Representation Learning for Heterogeneous Networks_诸神缄默不语的博客-CSDN博客)
- (KDD) Meta-Path Guided Embedding for Similarity Search in Large-Scale Heterogeneous Information Networks:提出ESim模型,根据预定义的权重从抽样出的metapaths实例中捕获节点语义
- (CIKM) HIN2Vec: Explore Meta-paths in Heterogeneous Information Networks for Representation Learning
- (WSDM) Embedding of Embedding (EOE): Joint Embedding for Coupled Heterogeneous Networks
2015年
2014年
- (KDD) Mining heterogeneous information networks: a structural analysis approach
- (CIKM) Meta-Path-Based Ranking with Pseudo Relevance Feedback on Heterogeneous Graph for Citation Recommendation
2013年
博弈论:
- (PLoS ONE) Evolution of Cooperation in a Heterogeneous Graph: Fixation Probabilities under Weak Selection
2011年
节点分类:
- (World Wide Web) Graffiti: graph-based classification in heterogeneous networks
节点相似性:
- (Proceedings of the VLDB Endowment) PathSim: Meta Path-Based Top-K Similarity Search in Heterogeneous Information Networks
2010年
通用节点嵌入:
- (ECML PKDD) Graph Regularized Transductive Classification on Heterogeneous Information Networks
- Optimal Embedding of Heterogeneous Graph Data with Edge Crossing Constraints