Comprehensive resource on graph representation learning (GRL), exploring fundamental principles, advanced methodologies, and case studies
A Complete Guide to Graph Representation Learning with Case Studies provides a concise understanding of the subject of graph representation learning (GRL), a rapidly advancing field in the domain of machine learning. The book explores basic concepts to state-of-the-art techniques, enabling readers to progress from a fundamental understanding of the approach to mastering its application. The authors also cover the topics of graph embedding methods, graph neural network (GNN) -based approaches, and the latest trends in GRL such as deep learning, transfer learning, graph pooling, alignment, and matching, and graph machine learning.
The book includes examples of applications of graph learning methods with real-world case studies in which the covered methods can be utilized. It also includes innovative solutions to graph machine learning problems such as node classification, link prediction, and unsupervised learning, and discusses neighborhood overlap visualization techniques and overlapping neighborhoods in heterogeneous graphs. Finally, the book provides an overview of open and ongoing research directions and student projects, providing a glimpse into potential avenues for future work.
The book also includes information on:
- Node-level features such as node degree, node centrality, closeness, betweenness, eigenvector, page rank centrality, clustering coefficient, closed triangles, egograph, and motifs
- Neighborhood sampling techniques such as breadth-first sampling, depth-first sampling, snowball sampling, random walk, shallow walk, edge sampling, link-based sampling, and metapath-based sampling
- Deep learning models including Graph Autoencoder (GAE), Variational Graph Encoder (VGAE), and Graph Attention Network (GAN)
- Graph alignment and matching, covering subgraph matching and embedding for matching
A Complete Guide to Graph Representation Learning with Case Studies is a thorough and up-to-date reference on the subject for engineers and researchers in data science and machine learning as well as graduate students in related programs of study.
Table of Contents:
List of Figures xxi
List of Tables xxvii
About the Book xxix
About the Authors xxxi
Preface xxxiii
List of Abbreviation xxxv
Part I Foundation Learning 1
1 Introduction to Graph and Graph Representation Learning 3
1.1 Introduction 3
1.2 What Is a Graph? 3
1.3 Importance of Graph 5
1.4 Types of Graphs 6
1.5 Overview of Graph Representation Learning (GRL) 13
1.6 Overview of Graph Connectivity 15
1.7 Foundation on Graph Neighborhood 17
1.8 Applications of Graph 23
1.9 Case Studies of GRL 25
1.10 Conclusion 31
2 Fundamental Concepts of Graph Structure 33
2.1 Introduction to Graph Structures 33
2.2 Node-Level Features 35
2.3 Structural-Level Features 46Contents vii
2.4 Graph-Level Features 52
2.5 Graph-Based Representation Techniques 56
2.6 Graph Representation Matrix 66
2.7 Conclusion 72
3 Overlapping Neighborhood in Graph 73
3.1 Definition and Its Importance 73
3.2 Methods for Detecting Neighborhood Overlap 75
3.3 Neighborhood Overlap Visualization Techniques 81
3.4 Overlapping Neighborhoods in Heterogeneous Graphs 84
3.5 Case Studies on Neighborhood Overlap Detection in Real-World Scenarios 92
3.6 Conclusion 95
Part II Core Graph Representation Learning 97
4 Graph Machine Learning 99
4.1 Introduction to GML 99Contents ix
4.2 Types of Tasks in GML 101
4.3 Conclusion 121
5 Graph Sampling 123
5.1 Introduction 123
5.2 Types of Graph Sampling 125
5.3 Case Studies 149
5.4 Conclusion 152
6 Graph Pooling 153
6.1 Introduction to Graph Pooling 153x Contents
6.2 Types of Graph Pooling 155
6.3 Case Study on Graph Pooling 178
6.4 Conclusion 179
7 Graph Neural Networks and Deep Representation Learning 181
7.1 Foundation of Deep Representation Learning on Graphs 181
7.2 Introduction to GNNs 185
7.3 Fundamentals of GCNs 189
7.4 Types of GNNs 194
7.5 Applications of Deep Learning on Graphs 207
7.6 Conclusion 208
8 Advanced Graph Neural Networks 211
8.1 Spatiotemporal GNNs 211
8.2 Dynamic GNNs 216
8.3 Hypergraph Neural Networks 223
8.4 Unsupervised Deep Learning Models 224
8.5 Conclusion 231
Part III Advanced Analysis and Techniques 233
9 Graph Alignment and Matching 235
9.1 Introduction to Graph Alignment and Matching 235
9.2 Definition and Purpose of Graph Alignment and Matching 235
9.3 Example for Graph Matching and Alignment 236
9.4 Types of Graph Alignment and Matching 237
9.5 Mathematical Approaches for Graph Alignment and Matching 241
9.6 Case Studies on Graph Alignment and Matching 246
9.7 Conclusion 247
10 Neighborhood Reconstruction Methods 249
10.1 Introduction 249
10.2 Neighborhood Reconstruction Techniques 251
10.3 Neighborhood Reconstruction Methods 262
10.4 Applications of Encoder–Decoder in Graph Learning 265
10.5 Case Study on Neighborhood Reconstruction 266
10.6 Conclusion 267
11 Transfer Learning on Graph 269
11.1 Introduction 269
11.2 Overview of TGL 270
11.3 Types of TGL 273
11.4 Key Techniques in TGL 284
11.5 Technique Comparisons 296
11.6 Case Studies 297
11.7 Conclusion 299
Part IV Emerging Trends with Case Studies 301
12 Graph Contrastive Learning 303
12.1 Introduction to Self-Supervised Learning 303
12.2 Comparison with Supervised Graph Learning 311
12.3 Fundamentals of GCL 312
12.4 GCL Frameworks 316
12.5 Conclusion 318
13 Multimodal Graph Representation Learning 321
13.1 Introduction to Multimodal GRL 321
13.2 Types of Modalities in Graph Data 322
13.3 Data Fusion Techniques in Multimodal GRL 325
13.4 Applications 331
13.5 Conclusion 334
14 Demystifying Graph Embeddings and Industrial Applications 337
14.1 Introduction 337
14.2 The Importance of Graph Representation 337
14.3 Making Sense Out of Graph-Structured Data 338
14.4 Delineating AI Model Engineering Steps 339
14.5 Graph Embeddings 340
14.6 Edge Embedding Techniques 348
14.7 Applications of Graph Embeddings 354
14.8 Conclusion 356
15 Knowledge Graph Foundation, Techniques, and Its Case Studies 359
15.1 Introduction to KG 359
15.2 Construction of a KG 365
15.3 Techniques for Building a KG 366
15.4 Representation and Storage 371Contents xvii
15.5 Case Studies of KG 374
15.6 Conclusion 383
16 Graph Representation Learning in Wireless Communication and Tourist Movement Analysis 385
16.1 A Graph-Theoretic Framework for Analyzing Tourist Flows from Social Media Data 385
16.2 Knowledge-Driven Graph Learning for Next-Generation Wireless Networks 390
16.3 Applications of GRL in Wireless Networks and Urban Tourism 395
16.4 Other Areas of GRL Applications 397
16.5 Conclusion 401
References 401
Index 403
About the Author :
E. CHANDRA BLESSIE, PhD, is Dean of Innovation with the School of Innovation at the KG College of Arts and Science Coimbatore Institute of Technology, Coimbatore, Tamil Nadu, India.
PETHURU RAJ CHELLIAH, PhD, is Vice President and Chief Architect of the Edge AI Division of Reliance Jio Platforms Ltd. in Bangalore, India.
B. SUNDARAVADIVAZHAGAN, PhD, is a Professor with the College of Computing and Information Sciences at the University of Technology and Applied Sciences Al Mussanah, Oman.
