Filters








57,660 Hits in 3.7 sec

CGMN: A Contrastive Graph Matching Network for Self-Supervised Graph Similarity Learning [article]

Di Jin, Luzhi Wang, Yizhen Zheng, Xiang Li, Fei Jiang, Wei Lin, Shirui Pan
2022 arXiv   pre-print
Then, we employ two strategies, namely cross-view interaction and cross-graph interaction, for effective node representation learning.  ...  In addition, existing unsupervised graph similarity learning methods are mainly clustering-based, which ignores the valuable information embodied in graph pairs.  ...  Both the view-and graph-level embedding are concatenated to the node embedding, which is used in building contrastiveness between two views to learn node embeddings of the graph.  ... 
arXiv:2205.15083v1 fatcat:a37v27nctjhchcbqmq2fija7qi

TCL: Transformer-based Dynamic Graph Modelling via Contrastive Learning [article]

Lu Wang, Xiaofu Chang, Shuang Li, Yunfei Chu, Hui Li, Wei Zhang, Xiaofeng He, Le Song, Jingren Zhou, Hongxia Yang
2021 arXiv   pre-print
In this paper, we propose a novel graph neural network approach, called TCL, which deals with the dynamically-evolving graph in a continuous-time fashion and enables effective dynamic node representation  ...  Secondly, on top of the proposed graph transformer, we introduce a two-stream encoder that separately extracts representations from temporal neighborhoods associated with the two interaction nodes and  ...  In this way, we can learn the underlying latent representations that the interaction nodes have in common.  ... 
arXiv:2105.07944v1 fatcat:4jn3wgiuonbqhnmlookwuo7p34

ConTIG: Continuous Representation Learning on Temporal Interaction Graphs [article]

Xu Yan, Xiaoliang Fan, Peizhen Yang, Zonghan Wu, Shirui Pan, Longbiao Chen, Yu Zang, Cheng Wang
2021 arXiv   pre-print
Representation learning on temporal interaction graphs (TIG) is to model complex networks with the dynamic evolution of interactions arising in a broad spectrum of problems.  ...  In the first update module, we employ a continuous inference block to learn the nodes' state trajectories by learning from time-adjacent interaction patterns between node pairs using ordinary differential  ...  Learning dynamic node embedding trajectories is extremely challenging due to the complex non-linear dynamic in temporal interaction graphs (TIG).  ... 
arXiv:2110.06088v1 fatcat:3zrmlzmmsrd6zkfz24culky2fa

GTEA: Representation Learning for Temporal Interaction Graphs via Edge Aggregation [article]

Yiming Li, Da Sun Handason Tam, Siyue Xie, Xiaxin Liu, Qiu Fang Ying, Wing Cheong Lau, Dah Ming Chiu, Shou Zhi Chen
2020 arXiv   pre-print
By capturing temporal interactive dynamics together with multi-dimensional node and edge attributes in a network, GTEA can learn fine-grained representations for a temporal interaction graph to enable  ...  The sequence model generates edge embeddings to encode temporal interaction patterns between each pair of nodes, while the GNN-based backbone learns the topological dependencies and relationships among  ...  Instead of learning from graph snapshots, JODIE (Kumar, Zhang, and Leskovec 2019) uses two recurrent neural networks to update node embeddings in each interaction.  ... 
arXiv:2009.05266v2 fatcat:catsc43x5fdozd4yfe2yfjmh3q

CGINet: graph convolutional network-based model for identifying chemical-gene interaction in an integrated multi-relational graph

Wei Wang, Xi Yang, Chengkun Wu, Canqun Yang
2020 BMC Bioinformatics  
We investigate two different perspectives on learning node embeddings.  ...  One is to view the graph as a whole, and the other is to adopt a subgraph view that initial node embeddings are learned from the binary association subgraphs and then transferred to the multi-interaction  ...  subgraph for final node embeddings learning.  ... 
doi:10.1186/s12859-020-03899-3 pmid:33243142 fatcat:2hhfej33hbgkxmg67bd5tjdnv4

GraphMS: Drug Target Prediction Using Graph Representation Learning with Substructures

Shicheng Cheng, Liang Zhang, Bo Jin, Qiang Zhang, Xinjiang Lu, Mao You, Xueqing Tian
2021 Applied Sciences  
We merged heterogeneous graph information and obtained effective node information and substructure information based on mutual information in graph embeddings.  ...  And the mutual information between the node–level and graph–level representations contributes most in a relatively dense network.  ...  In traditional graph embedding learning, nodes are adjacent to each other in the input diagram, and embedded represents are similar.  ... 
doi:10.3390/app11073239 fatcat:sshph6j5f5dx5cb4jev7f7ggye

Multilevel Graph Matching Networks for Deep Graph Similarity Learning [article]

Xiang Ling, Lingfei Wu, Saizhuo Wang, Tengfei Ma, Fangli Xu, Alex X. Liu, Chunming Wu, Shouling Ji
2021 arXiv   pre-print
In particular, the proposed MGMN consists of a node-graph matching network for effectively learning cross-level interactions between each node of one graph and the other whole graph, and a siamese graph  ...  Recent work on graph similarity learning has considered either global-level graph-graph interactions or low-level node-node interactions, however ignoring the rich cross-level interactions (e.g., between  ...  In particular, MGMN consists of a novel node-graph matching network (NGMN) for effectively learning cross-level interaction features by comparing each contextual node embedding of one graph against the  ... 
arXiv:2007.04395v4 fatcat:kp4u3tzdgfb33kvy2paxeksjcy

Localized Graph Collaborative Filtering [article]

Yiqi Wang, Chaozhuo Li, Mingzheng Li, Wei Jin, Yuming Liu, Hao Sun, Xing Xie, Jiliang Tang
2022 arXiv   pre-print
User-item interactions in recommendations can be naturally de-noted as a user-item bipartite graph.  ...  One key advantage of LGCF is that it does not need to learn embeddings for each user and item, which is challenging in sparse scenarios.  ...  The key advantage of GNN-based CF methods is to learn both the embeddings and a GNN model by explicitly capturing the structural information in the user-item interaction graph.  ... 
arXiv:2108.04475v2 fatcat:tmbrvj5ofzcvpc2i4ke7npce4i

Learning graph representations of biochemical networks and its application to enzymatic link prediction [article]

Julie Jiang, Li-Ping Liu, Soha Hassoun
2020 arXiv   pre-print
We show that ELP achieves high AUC when learning node embeddings using both graph connectivity and node attributes.  ...  We explore both transductive (test nodes included in the training graph) and inductive (test nodes not part of the training graph) learning models.  ...  The ELP method ELP has two steps ( Figure 1 ): (A) learning embedding vectors of graph nodes, and (B) predicting interaction between a pair of nodes from their embedding vectors.  ... 
arXiv:2002.03410v1 fatcat:i7waobftvfevreenohggc5dvdm

Modeling Pharmacological Effects with Multi-Relation Unsupervised Graph Embedding [article]

Dehua Chen, Amir Jalilifard, Adriano Veloso, Nivio Ziviani
2020 arXiv   pre-print
In this paper, we present a method based on a multi-relation unsupervised graph embedding model that learns latent representations for drugs and diseases so that the distance between these representations  ...  Compared with existing unsupervised graph embedding methods our method shows superior prediction performance in terms of area under the ROC curve, and we present examples of repositioning opportunities  ...  The graph also contains protein-protein interactions in order to increase connectivity and information propagation while learning node representations.  ... 
arXiv:2004.14842v2 fatcat:xsyhjkho7jesndalsvh47cfye4

Interest-aware Message-Passing GCN for Recommendation [article]

Fan Liu, Zhiyong Cheng, Lei Zhu, Zan Gao, Liqiang Nie
2021 arXiv   pre-print
neighboring users with no common interests of a user can be also involved in the user's embedding learning in the graph convolution operation.  ...  Graph Convolution Networks (GCNs) manifest great potential in recommendation.  ...  In IMP-GCN, the embedding of a node learned in a subgraph only contributes to the embedding learning of other nodes in this subgraph.  ... 
arXiv:2102.10044v2 fatcat:q2npm2v5bnasrhvz6z56fsvy24

APAN: Asynchronous Propagation Attention Network for Real-time Temporal Graph Embedding [article]

Xuhong Wang, Ding Lyu, Mengjian Li, Yang Xia, Qi Yang, Xinwen Wang, Xinguang Wang, Ping Cui, Yupu Yang, Bowen Sun, Zhenyu Guo
2021 arXiv   pre-print
Therefore, we propose Asynchronous Propagation Attention Network, an asynchronous continuous time dynamic graph algorithm for real-time temporal graph embedding.  ...  Limited by the time complexity of querying k-hop neighbors in a graph database, most graph algorithms cannot be deployed online and execute millisecond-level inference.  ...  to expand GCN into inductive learning which also works on unknown nodes in graphs.  ... 
arXiv:2011.11545v4 fatcat:4bzifh5vfrdplail233b2m2hni

Graph Embedding on Biomedical Networks: Methods, Applications, and Evaluations [article]

Xiang Yue, Zhen Wang, Jingong Huang, Srinivasan Parthasarathy, Soheil Moosavinasab, Yungui Huang, Simon M. Lin, Wen Zhang, Ping Zhang, Huan Sun
2019 arXiv   pre-print
Graph embedding learning which aims to automatically learn low-dimensional node representations has drawn increasing attention in recent years.  ...  , protein-protein interaction prediction, and two node classification tasks: medical term semantic type classification, protein function prediction.  ...  DeepWalk is then performed on the multilayer graph to learn node representations in which nodes with high structural similarity are close to each other in the embedding space.  ... 
arXiv:1906.05017v2 fatcat:7k6vrdkwybdu7pikmxa3xrnowm

Feature Interaction-aware Graph Neural Networks [article]

Kaize Ding, Yichuan Li, Jundong Li, Chenghao Liu, Huan Liu
2020 arXiv   pre-print
In this paper, we propose Feature Interaction-aware Graph Neural Networks (FI-GNNs), a plug-and-play GNN framework for learning node representations encoded with informative feature interactions.  ...  Specifically, the proposed framework is able to highlight informative feature interactions in a personalized manner and further learn highly expressive node representations on feature-sparse graphs.  ...  In this way, we are able to learn feature interaction-aware node representations on such feature-sparse graphs.  ... 
arXiv:1908.07110v2 fatcat:fk76u5vxorfljom5s3ivwhk6ku

Curriculum Pre-Training Heterogeneous Subgraph Transformer for Top-N Recommendation [article]

Hui Wang, Kun Zhou, Wayne Xin Zhao, Jingyuan Wang, Ji-Rong Wen
2021 arXiv   pre-print
Besides, we design a curriculum pre-training strategy to provide an elementary-to-advanced learning process, by which we smoothly transfer basic semantics in HIN for modeling user-item interaction relation  ...  Due to the flexibility in modelling data heterogeneity, heterogeneous information network (HIN) has been adopted to characterize complex and heterogeneous auxiliary data in top-N recommender systems, called  ...  Besides node ID embeddings, we introduce node type embedding, slot embedding and precursor embedding to preserve the semantics of interaction-specific heterogeneous subgraphs in multi-slot sequence representations  ... 
arXiv:2106.06722v1 fatcat:5sxsricufveodho3fblxxrhb2a
« Previous Showing results 1 — 15 out of 57,660 results