This book provides a comprehensive introduction to the foundations and frontiers of graph neural networks. In addition, the book introduces the basic concepts and definitions in graph representation learning and discusses the development of advanced graph representation learning methods with a focus on graph neural networks. The book providers researchers and practitioners with an understanding of the fundamental issues as well as a launch point for discussing the latest trends in the science. The authors emphasize several frontier aspects of graph neural networks and utilize graph data to describe pairwise relations for real-world data from many different domains, including social science, chemistry, and biology. Several frontiers of graph neural networks are introduced, which enable readers to acquire the needed techniques of advances in graph neural networks via theoretical models and real-world applications.
Author(s): Chuan Shi, Xiao Wang, Cheng Yang
Series: Synthesis Lectures on Data Mining and Knowledge Discovery
Publisher: Springer
Year: 2022
Language: English
Pages: 206
City: Cham
535226_1_En_OFC
535226_1_En_BookFrontmatter_OnlinePDF
Foreword
Preface
Contents
535226_1_En_1_Chapter_OnlinePDF
1 Introduction
1.1 Basic Concepts
1.1.1 Graph Definitions and Properties
1.1.2 Complex Graphs
1.1.3 Computational Tasks on Graphs
1.2 Development of Graph Neural Network
1.2.1 History of Graph Representation Learning
1.2.2 Frontier of Graph Neural Networks
1.3 Organization of the Book
535226_1_En_2_Chapter_OnlinePDF
2 Fundamental Graph Neural Networks
2.1 Introduction
2.2 Graph Convolutional Network
2.2.1 Overview
2.2.2 The GCN Method
2.3 Inductive Graph Convolution Network
2.3.1 Overview
2.3.2 The GraphSAGE Method
2.4 Graph Attention Network
2.4.1 Overview
2.4.2 The GAT Method
2.5 Heterogeneous Graph Attention Network
2.5.1 Overview
2.5.2 The HAN Method
535226_1_En_3_Chapter_OnlinePDF
3 Homogeneous Graph Neural Networks
3.1 Introduction
3.2 Adaptive Multi-channel Graph Convolutional Networks
3.2.1 Overview
3.2.2 Investigation
3.2.3 The AM-GCN Method
3.2.4 Experiments
3.3 Beyond Low-Frequency Information in Graph Convolutional Networks
3.3.1 Overview
3.3.2 Investigation
3.3.3 The FAGCN Method
3.3.4 Experiments
3.4 Graph Structure Estimation Neural Networks
3.4.1 Overview
3.4.2 The GEN Method
3.4.3 Experiments
3.5 Interpreting and Unifying GNNs with An Optimization Framework
3.5.1 Overview
3.5.2 Preliminary
3.5.3 The GNN-LF/HF Method
3.5.4 Experiments
3.6 Conclusion
3.7 Further Reading
535226_1_En_4_Chapter_OnlinePDF
4 Heterogeneous Graph Neural Networks
4.1 Introduction
4.2 Heterogeneous Graph Propagation Network
4.2.1 Overview
4.2.2 The HPN Method
4.2.3 Experiments
4.3 Heterogeneous Graph Neural Network with Distance Encoding
4.3.1 Overview
4.3.2 The DHN Method
4.3.3 Experiments
4.4 Self-supervised HGNN with Co-contrastive Learning
4.4.1 Overview
4.4.2 The HeCo Method
4.4.3 Experiments
4.5 Conclusion
4.6 Further Reading
535226_1_En_5_Chapter_OnlinePDF
5 Dynamic Graph Neural Networks
5.1 Introduction
5.2 Micro- and Macro-dynamics
5.2.1 Overview
5.2.2 The M2DNE Method
5.2.3 Experiments
5.3 Heterogeneous Hawkes Process
5.3.1 Overview
5.3.2 The HPGE Method
5.3.3 Experiments
5.4 Dynamic Meta-Path
5.4.1 Overview
5.4.2 The DyMGNN Method
5.4.3 Experiments
5.5 Conclusion
5.6 Further Reading
535226_1_En_6_Chapter_OnlinePDF
6 Hyperbolic Graph Neural Networks
6.1 Introduction
6.2 Hyperbolic Graph Attention Network
6.2.1 Overview
6.2.2 The HAT Method
6.2.3 Experiments
6.3 Lorentzian Graph Convolutional Network
6.3.1 Overview
6.3.2 The LGCN Method
6.3.3 Experiments
6.4 Hyperbolic Heterogeneous Graph Representation
6.4.1 Overview
6.4.2 The HHNE Method
6.4.3 Experiments
6.5 Conclusion
6.6 Further Reading
535226_1_En_7_Chapter_OnlinePDF
7 Distilling Graph Neural Networks
7.1 Introduction
7.2 Prior-Enhanced Knowledge Distillation for GNNs
7.2.1 Overview
7.2.2 The CPF Method
7.2.3 Experiments
7.3 Temperature-Adaptive Knowledge Distillation for GNNs
7.3.1 Overview
7.3.2 The LTD Method
7.3.3 Experiments
7.4 Data-Free Adversarial Knowledge Distillation for GNNs
7.4.1 Overview
7.4.2 The DFAD-GNN Method
7.4.3 Experiments
7.5 Conclusion
7.6 Further Reading
535226_1_En_8_Chapter_OnlinePDF
8 Platforms and Practice of Graph Neural Networks
8.1 Introduction
8.2 Foundation
8.2.1 Deep Learning Platforms
8.2.2 Platforms of Graph Neural Networks
8.2.3 GammaGL
8.3 Practice of Graph Neural Networks on GammaGL
8.3.1 Create Your Own Graph
8.3.2 Create Message-Passing Network
8.3.3 Advanced Mini-Batching
8.3.4 Practice of GIN
8.3.5 Practice of GraphSAGE
8.3.6 Practice of HAN
8.4 Conclusion
535226_1_En_9_Chapter_OnlinePDF
9 Future Direction and Conclusion
9.1 Future Direction
9.1.1 Self-supervised Learning on Graphs
9.1.2 Robustness
9.1.3 Explainability
9.1.4 Fairness
9.1.5 Biochemistry
9.1.6 Physics
9.2 Conclusion
535226_1_En_BookBackmatter_OnlinePDF
References
