metapath2vec: Scalable Representation Learning for Heterogeneous Networks

By Yuxiao Dong, Nitesh V. Chawla, Ananthram Swami. 2017

2,469 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistical and Nonlinear Physics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistical and Nonlinear Physics

Citations in the corpus

Representation Learning on Graphs: Methods and Applications (2017, background): “7 Recently, Dong et al. [19] also proposed a strategy for sampling random walks from heterogeneous graphs, where the random walks are restricted to only transition between particular types of nodes.”
A Comprehensive Survey of Graph Embedding: Problems, Techniques, and Applications (2017, background): “[46] design meta-path-based randomwalks to deal with heterogeneous graphs and a heterogeneous SkipGram which maximizes the probability of having the hetegeneous context for a given node.”
Network Embedding as Matrix Factorization: Unifying DeepWalk, LINE, PTE, and node2vec (2017, background): “…Deep-Walk [28], LINE [35], a growing number of literature has paid the attention to this area, for example, heterogeneous network embedding [7, 11, 17, 34], direct network embedding [27], semi-supervised network embedding [18, 42, 49], network embedding with rich vertex attributes [41, 47],…”
Heterogeneous Graph Attention Network (2019, methodology): “paths. To achieve the best performance, ESim needs to conduct grid search to find the optimal weights of hmeta-paths. It is pretty hard to find the optimal combination for specific task. Metapath2vec [7] designs a meta-path based random walk and utilizes skip-gram to perform heterogeneous graph embe”
Heterogeneous Information Network Embedding for Recommendation (2017, methodology): “[58] define the neighbor of nodes via meta-path and learn the heterogeneous embedding by skip-gramwith negative sampling.”
Heterogeneous Graph Structure Learning for Graph Neural Networks (2021, methodology): “…2017), GAT (Velick-ovic et al. 2018), and GraphSAGE (Hamilton, Ying, and Leskovec 2017), four heterogeneous graph embedding methods, i.e., MP2Vec (Dong, Chawla, and Swami 2017), HAN (Wang et al. 2019b), HeGAN (Hu, Fang, and Shi 2019), and GTN (Yun et al. 2019), and three graph structure learning…”
A Survey on Knowledge Graph-Based Recommender Systems (2020, methodology): “In the ﬁrst phase, KTGAN learns the knowledge embedding v kj for movie v j by incorporating the Metapath2Vec model [105] on the movie’s KG, and the tag embedding v tj with the Word2Vec model [106] on movie’s attributes.”
Propagation Model Search for Graph Neural Networks (2020, methodology): “We compare against unsupervised methods including DeepWalk (Per-ozzi, Al-Rfou, and Skiena 2014) and metapath2vec (Dong, Chawla, and Swami 2017), homogeneous GNNs including GCN (Kipf and Welling 2017) and GAT (Veliˇckovi´c et al. 2018), and heterogeneous GNNs including HAN (Wang et al. 2019) and GTN…”
Representation Learning for Dynamic Graphs: A Survey (2019, methodology): “A meta-path can be considered as a sequence of relations in R. Dong et al. (2017) propose metapath2vec where each random walk is constrained to conform to a meta-path r1, r2, . . . , rk by starting randomly at a node v1 whose type is compatible with the source type of r1.”
Open Graph Benchmark: Datasets for Machine Learning on Graphs (2020, methodology): “[26] Yuxiao Dong, Nitesh V Chawla, and Ananthram Swami.”
A Semi-Supervised Graph Attentive Network for Financial Fraud Detection (2019, background): “Then some further works like Metapath2vec [22] , HNE [23] are proposed to consider both the network topology”
Network Representation Learning: A Survey (2017, neutral): “Recent studies by [16], [113], [114], [115], [116], [117], [118], [119],”
Leveraging Domain Context for Question Answering Over Knowledge Graph (2019, methodology): “We adopt the meta-path random walk [10] to capture the context of the targeted entities in the knowledge graph.”