heterogeneous graph transformer
play

Heterogeneous Graph Transformer WWW20 1 Author Second-year CS - PowerPoint PPT Presentation

Heterogeneous Graph Transformer WWW20 1 Author Second-year CS Ph.D student, advised by Prof. Yizhou Sun bachelor degree in Peking University, advised by Prof. Xuanzhe Liu. WSDM 2018, WWW 2019, Best Paper Award, ICLR 2019 Workshop,


  1. Heterogeneous Graph Transformer WWW20 1

  2. Author • Second-year CS Ph.D student, advised by Prof. Yizhou Sun • bachelor degree in Peking University, advised by Prof. Xuanzhe Liu. WSDM 2018, WWW 2019, Best Paper Award, ICLR 2019 Workshop, ACL 2019, WWW 2020 2

  3. Background • General GNN Framework s Aggr s t s 3

  4. Background • Graph Attention Network s Aggr s t s 4

  5. Background • Relational graph convolutional networks (R-GCN) ($) 𝑋 # 𝑤 " ℎ " !" 𝑠 ! 𝑠 " # ℎ " 𝑤 " 𝑤 ! 𝑠 ! # 𝑤 " ℎ " ($) 𝑋 !" 5

  6. Background • Node classification 西虹 市首 喜剧 富 沈 腾 羞羞 ? 的铁 拳 6

  7. Heterogeneous Information Networks (HIN) From Heterogeneous Graph Neural Network and its Applications in E-Commerce Prof. Chuan Shi 7

  8. OAG Graph 8

  9. OAG Graph 9

  10. Tasks • Node Classification • Paper-Field prediction • Paper–Field (L1) p1 • Paper–Field (L2) 哈工 大 • Paper-Venue prediction zwn p2 • Link prediction 上交 zwn • Author Disambiguation tasks p3 10

  11. Heterogeneous Graph Directed graph Type mapping functions 𝑓 = (𝐼𝐻𝑈, 𝐼𝐵𝑂) 𝑤 =< Heterogeneous Graph Transformer > 𝜐 𝑤 = 𝑞𝑏𝑞𝑓𝑠 ∅ 𝑓 = 𝑑𝑗𝑢𝑓𝑒 11

  12. Meta Relation 𝑓 = (𝐼𝐻𝑈, 𝐼𝐵𝑂) < 𝜐 𝑡 , ∅ 𝑓 , 𝜐 𝑢 > =< 𝑞𝑏𝑞𝑓𝑠, 𝑑𝑗𝑢𝑓𝑒, 𝑞𝑏𝑞𝑓𝑠 > 12

  13. Model • Heterogeneous Mutual Attention • Heterogeneous Message Passing • Target-Specific Aggregation 13

  14. Heterogeneous Mutual Attention s Aggr s t s 14

  15. Heterogeneous Mutual Attention e e d d g g e e 1 1 edge1 edge1 type1 s type1 s s t s t type1 type2 type1 type1 edge1 edge1 edge2 edge2 s s edge2 edge2 type2 type2 15

  16. Heterogeneous Message Passing edge1 type1 s s t type1 edge1 edge2 s type2 16

  17. Target-Specific Aggregation edge1 type1 s s t type1 edge1 type1 edge2 s type2 17

  18. Overall Architecture 18

  19. Dynamic Heterogeneous Graph 𝑤 = 𝐼𝐻𝑈 𝑤 = 𝐼𝐵𝑂 𝑓 = (𝐼𝐻𝑈, 𝑋𝑋𝑋) 𝑓 = (𝐼𝐻𝑈, 𝑋𝑋𝑋) 𝑋𝑋𝑋 2020 𝑋𝑋𝑋 2019 𝑓 = (𝐼𝐻𝑈, 𝑋𝑋𝑋) timestamp 2020 𝑓 = (𝐼𝐵𝑂, 𝑋𝑋𝑋) timestamp 2019 𝑤 = 𝐼𝐻𝑈 timestamp 2020 𝑤 = 𝐼𝐵𝑂 timestamp 2019 𝑤 = 𝑋𝑋𝑋 timestamp 2020 𝑤 = 𝑋𝑋𝑋 timestamp 2019 19

  20. Relative Temporal Encoding 2 0 Transformer Relative Temporal Encoding 20

  21. Relative Temporal Encoding 21

  22. Overall Architecture 22

  23. HGSampling • keep a similar number of nodes and edges for each type • keep the sampled sub-graph dense to minimize the information loss and reduce the sample variance. 23

  24. Baselines • GCN • GAT • R-GCN • HetGNN (KDD19 Heterogeneous Graph Neural Network) • HAN (WWW19 Heterogeneous Graph Attention Network) 24

  25. Input Features • Paper • pre-trained XLNet to get the representation of each word in its title. • Average them weighted by each word’s attention to get the title representation for each paper. • Author • average of his/her published papers’ representations • Field, venue, and institute • metapath2vec 25

  26. Results 26

  27. Visualize Meta Relation Attention 27

  28. Papers 28

  29. Thanks! 29

Recommend


More recommend