PART I : FOUNDATION LEARNING
CHAPTER 1 : INTRODUCTION TO GRAPH AND GRAPH REPRESENTATION LEARNING (GRL)
1. Introduction
1.1 What is a Graph?
1.2 Importance of Graph
1.3 Types of Graph
1.3.1 Directed Graph
1.3.2 Undirected Graph
1.3.3 Weighted Graph
1.3.4 Multigraph
1.3.5 Complete Graph
1.3.6 Planar and Bipartite Graph
1.3.7 Hypergraph
1.3.8 Hierarchical Graphs
1.3.9 Homogeneous and Heterogeneous Graphs
1.4 Overview of Graph Representation Learning
1.4.1 Benefits of Graph Representation Learning (GRL)
1.5 Overview of Graph Connectivity
1.5.1 Its Definition
1.5.2. Importance of Graph Connectivity
1.5.3. Role of Graph Connectivity
1.6 Foundation on Graph Neighborhood
1.6.1 Introduction
1.6.2 Properties of Graph Neighborhoods
1.6.3 Types of Node/Graph Neighborhoods
1.6.3.1 Direct/Local Neighborhood (1-Hop Neighborhood)
1.6.3.2 K-Hop Neighborhood
1.6.3.3 Ego Neighborhood
1.6.3.4 Attention-based neighborhoods
1.6.3.5 Structural Neighborhood (Graphlets and Motifs)
1.6.3.6 Graphlets
1.7 Applications of Graph
1.7.1 Traffic Prediction in Transportation
1.7.2 Pattern recognition
1.7.3 Social Network Analysis
1.7.4 Application of Graphs in Biology
1.7.5 Application of graphs in chemistry
1.8 Case studies of Graph Representation Learning (GRL)
1.8.1 Case study in Social Network Analysis
1.8.2 Case Study in Fraud Detection for Financial Transactions
1.8.3 Case Study in Molecular Biology
1.9 Conclusion
References
CHAPTER 2 : FUNDAMENTAL CONCEPTS OF GRAPH STRUCTURES
2.1 Introduction to Graph Structures
2.1.1 Characteristics of Graph Structure
2.2 Node-Level features
2.2.1 Node Degree
2.2.2 Node Centrality Measures
2.3 Structural-Level features
2.3.1 Clustering Coefficient measure
2.3.2 Closed Triangles
2.3.3 Egographs
2.3.4 Motifs
2.4 Graph-Level Features
2.4.1 Structural Properties of Graphs
2.4.1.1 Graph Degree and Density
2.4.1.2 Graph Eccentricity and Modularity
2.5 Graph-Based Representation Techniques
2.5.1 Graph Clustering Coefficient
2.5.2 Bag of Nodes (BoN)
2.5.3 Weifeiler-Lehman Kernel
2.5.4 Graphlets and Path-Based Methods
2.6 Graph Representation Matrix
2.6.1 Types of Graph Representation Matrices
2.6.1.1 Adjacency Matrix
2.6.1.2 Incidence Matrix
2.6.1.3 Laplacian Matrix
2.6.1.4 Transition Matrix
2.7 Conclusion
References
CHAPTER 3 OVERLAPPING NEIGHBORHOOD IN GRAPH
3.1 Definition and its importance
3.2 Methods for Detecting Neighborhood overlap
3.2.1 Local Overlap Measures
3.2.1.1 Jaccard Similarity measure
3.2.1.2 Adamic-Adar Index
3.2.1.3 Resource Allocation Index
3.2.1.4 Global Overlap Measures
3.2.1.5 Katz Index
3.2.1.6 SimRank
3.2.1.7 PageRank-Based Similarity
3.3 Neighborhood Overlap Visualization Techniques
3.3.1 Venn Diagram
3.3.2 Node-Link Diagrams (Graph Representation)
3.3.3 Heatmaps
3.3.4 Edge Bundling
3.3.5 Graph Embedding & Projection (t-SNE, UMAP)
3.3.6 Ego Networks (Subgraph Extraction )
3.4 Overlapping Neighborhoods in Heterogeneous Graphs
3.4.1 Introduction
3.4.2 Methods for Detecting Overlapping Neighborhoods in Heterogeneous Graphs
A. Meta-Path Based Approaches
B. Clustering-based methods
3.5 Case Studies on neighborhood overlap detection in real-world scenarios
3.5.1 Application of Graph Neighborhood in Healthcare
3.5.2 Case studies on Social networks: overlapping user communities and friend
recommendations
3.6 Conclusion
References
PART II : CORE GRAPH REPRESENTATION LEARNING
CHAPTER 4 : GRAPH MACHINE LEARNING
4.1 Introduction to Graph Machine Learning (GML)
4.1.1 Key features of GML
4.1.2 GML Vs traditional ML
4.2 Types of Tasks in GML
4.2.1 Supervised Graph Machine Learning (GML) tasks
4.2.1.1 Node Classification
4.2.1.2 Link Prediction
4.2.1.3 Edge Classification
4.2.1.4 Graph / Subgraph classification
4.2.2 Unsupervised Graph Machine Leaning tasks
4.2.2.1 Introduction
4.2.2.2 Node Clustering
A. Spectral Clustering
B. Label Propagation Algorithm (LPA)
C. Louvain Method for Community Detection
4.2.3 Case studies and Applications of GML
Case study 1 : GML in Disease prediction
Case study 2 : Drug Discovery and Development
4.3 Conclusion
References
CHAPTER 5 : GRAPH SAMPLING
5.1 Introduction
5.1.1 Issues related to large and complex graphs
5.1.2 Definition with examples
5.1.3 Challenges of using sampling
5.2 Types of Graph Sampling
5.2.1 Node Sampling
5.2.1.1 Overview of Node Sampling
5.2.1.2 Types of Node Sampling
A. Uniform Node Sampling
B. Degree-based Node Sampling
C. Stratified Node Sampling
D. Snowball Node Sampling
E. Random Walk-based Sampling
5.2.2 Edge Sampling
5.2.2.1 Overview of Edge Sampling
5.2.2.2 Types of Edge Sampling
A. Uniform Edge Sampling
B. Weighted Edge Sampling
C. Temporal Edge Sampling
5.2.3 Neighborhood Sampling
5.2.3.1 Overview of Neighborhood Sampling
5.2.3.2 Types of Neighborhood Sampling
A. K-Hop Neighborhood Sampling
B. Random Walk Neighborhood Sampling
C. Layer-wise Neighborhood Sampling
5.2.4 Subgraph sampling
5.3 Case Studies
5.3.1 Case Study: Sampling Patient Networks for Disease Prediction
5.3.2 Case Study: Sampling in Learning Management System (LMS)
5.3.3 Case Study: Community-Based Sampling for Efficient Social Network Analysis
5.4 Conclusion
References
CHAPTER 6 : GRAPH POOLING
6.1 Introduction to Graph Pooling
6.1.1 What is Graph Pooling?
6.1.2 Why is Graph Pooling Important?
6.2 Types of Graph Pooling
6.2.1 Flat Pooling
6.2.1.1Methods of flat Pooling
A. Sum Pooling
B. Mean Pooling
C. Max Pooling
D. Attention Pooling
6.2.1.2 Applications of Flat Pooling in Graph-Level Tasks
6.2.2. Hierarchical Pooling
6.2.2.1 Types of Hierarchical Pooling
6.2.2.2Common techniques
A. Top-K Pooling
B. DiffPool (Differentiable Pooling)
C. SAGPool (Self-Attention Graph Pooling)
D. MinCutPool (Minimum Cut Pooling)
6.2.3 Comparison between Flat pooling and Hierarchical pooling
6.3 Case Study on Graph Pooling
6.4 Conclusion
References
CHAPTER 7 : GRAPH NEURAL NETWORKS AND DEEP REPRESENTATION LEARNING
7.1 Foundation of Deep Representation Learning on graphs
7.1.1 Basic concept
7.1.2 Challenges and importance of Deep Learning on Graphs
7.1.3 Approach of Deep Learning on Graph
7.1.4 Characteristics and Benefits of Deep Learning on Graph
7.2 Introduction to Graph Neural Networks (GNNs)
7.2.1 GNN Architecture Overview
7.2.2 Message Passing in GNNs
7.2.3 Case Study on Graph Neural Network (GNNs)
7.2.3.1 Case Study 1: Enhancing Credit Card Fraud Detection with CaT-GNN
7.2.3.2 Case Study: GNNs for Drug-Drug Interaction (DDI) Prediction in
Healthcare
7.3 Fundamentals of Graph Convolution Networks (GCNs)
7.3.1 Core concepts and motivation for GCNs
7.3.2 Workflow process of Graph Convolution Network (GCN)
7.4 Types of Graph Neural Networks
7.4.1 Graph Recurrent Neural Networks (GRNN)
7.4.1.1 Definition with illustration
7.4.1.2 Working process of GRNN
7.4.1.3 GRNN Architecture
7.4.1.4 Applications of GRNN
7.4.2 Graph Attention Networks (GAT)
7.4.2.1 Definition with illustration
7.4.2.2 Working Process of Graph Attention Networks (GAT)
7.4.2.3 Architecture of GAT
7.4.2.4 Application of GAT
7.5 Applications of Deep Learning on Graph
7.6 Conclusion
References
CHAPTER 8 : ADVANCED GRAPH NEURAL NETWORKS
8.1 Spatial-Temporal Graph Neural Networks (ST-GNNs)
8.1.1 Definition with illustration
8.1.2 Working Process of Spatial-Temporal Graph Neural Networks (ST-GNN)
8.1.3 Architecture of ST-GNNs
8.1.4 Application of ST-GNNs
8.2 Dynamic Graph Neural Networks
8.2.1 Work Process of Dynamic GNNs
8.2.2 Example: Dynamic Graph â" Social Interaction
8.2.3 Dynamic Graph Neural Network models
A. Temporal Graph Attention Network (TGAT)
B. Dynamic Graph Convolutional Network (DyGCN)
8.3 Hypergraph Neural Networks
8.3.1 Architecture of HGNN
8.3.2 Working Process of HGNN
8.4 Unsupervised Deep Learning Models
8.4.1 Graph Autoencoder (GAE)
8.4.1.1 GAE â" Workflow with its Architecture
8.4.1.2 GAE Algorithm
8.4.1.3 Applications of GAE
8.4.1.4 Advantages and Limitations of GAE
8.4.2 Variational Auto Encoder (VAE)
8.4.2.1 VGAE â" Workflow with its Architecture
8.4.2.2 Algorithm: Variational Graph Autoencoder (VGAE)
8.4.2.3 Advantages and Limitations of VGAE
8.5 Conclusion
References
PART III : ADVANCED ANALYSIS AND TECHNIQUES
CHAPTER 9 : GRAPH ALIGNMENT AND MATCHING
9.1 Introduction to Graph Alignment and Matching
9.2 Definition and purpose of graph alignment and matching.
9.3 Example for Graph Matching and Alignment
9.4. Types of Graph Alignment and Matching
9.4.1. Exact Graph Matching (Graph Isomorphism)
9.4.2. Inexact/Fuzzy Matching
9.4.3. Global Alignment
9.4.4. Local Alignment
9.5. Mathematical Approaches for Graph Alignment and Matching
9.5.1. Graph Embedding + Distance Minimization
9.5.2 Gromov-Wasserstein distance
9.5.3 Contrastive Learning for Graph Matching
9.5.4 Deep Graph Matching Networks (e.g., DGMC)
9.6. Case Studies on Graph Alignment and Matching
9.6.1 Graph Alignment and Matching in Healthcare
9.6.2. Matching Users Across Facebook and Twitter Accounts in Social Networks
9.7 Conclusion
References
CHAPTER 10 : NEIGHBORHOOD RECONSTRUCTION METHODS
10.1 Introduction
10.1.1 Introduction to Neighborhood Reconstruction in Graphs
10.1.2 Benefits of Neighborhood Reconstruction in Graphs
10.1.3 Importance of reconstructing graph structures
10.2 Neighborhood Reconstructions Techniques
10.2.1 Matrix Factorization
10.2.2 Embedding Models
10.2.3 Similarity Measures
10.3 Neighborhood reconstruction methods
10.3.1 An Encoder-Decoder method
10.4. Applications of Encoder-Decoder in Graph Learning
10.5 Case Study on Neighborhood reconstruction
10.5.1Case Study on Urban Road Network Reconstruction Using Neighborhood Inference
10.6 Conclusion
References
CHAPTER 11 : TRANSFER LEARNING ON GRAPH
11.1 Introduction
11.2 Overview of Transfer graph learning
11.2.1 Key features of TGL
11.2.2 Benefits of Transfer Graph Learning
11.2.3 The need of TGL and its necessity
11.2.4 TGL Vs GL
11.3 Types of transfer learning (TGL)
11.3.1 Inductive Transfer Graph Learning (ITGL)
11.3.1.1 Working Process of Inductive Transfer Graph Learning
11.3.1.2 Algorithm : GraphSAGE (Graph Sample and Aggregate)
11.3.1.3 Key Features and Benefits of GraphSAGE
11.3.2 Transductive Transfer Graph Learning (TTGL)
11.3.2.1 Working Process of Transductive Transfer Graph Learning
11.3.2.2 Algorithm : GCN-DA (Graph Convolutional Network with Domain Adaptation)
11.3.2.3 Key Features and Benefits of GCN-DA
11.3.3 Heterogeneous Transfer Learning (HTL)
11.3.3.1 Working Process of Heterogeneous Transfer Learning
11.3.3.2 Algorithm : Heterogeneous Feature Augmentation (HFA)
11.3.3.3 Key Features and Benefits of HFA
11.4 Key Techniques in Transfer Graph Learning (TGL)
11.4.1 Feature-Based Transfer Techniques
11.4.1.1 Transfer Component Analysis (TCA)
11.4.1.2 Domain Adversarial Neural Networks (DANN)
11.4.2 Structure-Based Transfer Techniques
11.4.2.1 Overview and its Definition
11.4.2.2 Algorithm for Structure-Based Transfer Learning
11.4.2.3 Flow process and its diagram
11.4.2.4 Meta-Path-Based Learning
11.5 Techniques Comparison
11.5.1 Inductive vs. Transductive Transfer Graph Learning
11.5.2 Feature-based transfer graph learning vs structure-based transfer graph learning
11.6 Case studies
11.6.1 Case Study: Enhancing Graph Anomaly Detection with Dual-Level Knowledge Distillation
11.6.2 Transfer Learning in Biological Networks using Geneformer
11.7 Conclusion
References
PART IV : EMERGING TRENDS WITH CASE STUDIES
CHAPTER 12 : GRAPH CONTRASTIVE LEARNING
12.1 Introduction to Self-Supervised Learning
12.1.1 Definition with example
12.1.2 Key Characteristics of Self-Supervised Learning (SSL)
12.1.3 Types of Self-Supervised Tasks
12.1.4 Application in Graph Domain
12.2 Comparison with Supervised Graph Learning
12.3 Fundamentals of Graph Contrastive Learning (GCL)
12.3.1 Motivation : Why Contrastive Learning for graphs?
12.3.2 Key Components and Mechanism
12.4 Graph Contrastive Learning Frameworks
12.5 Conclusion
References
CHAPTER 13 : MULTIMODEL GRAPH REPRESENTATION LEARNING
13.1 Introduction to Multimodal GRL
13.2 Types of Modalities in Graph data
13.2.1 Text Modalities
13.2.2 Image Modalities
13.2.3 Audio Modalities
13.3 Data Fusion Techniques in Multimodal GRL
13.3.1 What is Data Fusion?
13.3.2Data Fusion Techniques
13.4 Applications
13.4.1 Autonomous Vehicles
13.4.2 Agriculture â" Pest Detection Network
13.5 Conclusion
References
CHAPTER 14 : DEMYSTIFYING GRAPHS EMBEDDINGS AND INDUSTRIAL APPLICATIONS
14.1 Introduction
14.2 The Importance of Graph Representation
14.3 Making Sense out of Graph-structured Data
14.4 Delineating AI Model Engineering Steps
14.5 Graph Embeddings
14.5.1 The Need for Graph Embeddings
14.5.2 Knowledge Graph Embeddings
14.5.3 The Advantages of Graph Embeddings
14.5.4 Graph Embedding Applications
14.5.5 Types of Graph Embeddings
14.6 Edge embedding techniques
14. 7 Applications of Graph Embeddings
14.8 Conclusion
References
Chapter 15 : KNOWLEDGE GRAPH FOUNDATION, TECHNIQUES WITH ITS CASE STUDIES
15.1 Introduction to Knowledge Graph
15.1.1 History and evolution
15.1.2 Definition and core concepts
15.1.3 Difference between KGs and traditional databases
15.2 Construction of a Knowledge Graph
15.3 Techniques for Building a Knowledge Graph
15..3.1 Named Entity Recognition (NER)
15.3.2 Entity Linking and Disambiguation
15.3.3 Relation Extraction (RE)
15.3.4 Ontology Engineering
15.4 Representation and Storage
15.4.1 Introduction to Representation and Storage of Knowledge Graphs
15.4.2 RDF Triples and SPARQL
15.5 Case Studies of Knowledge Graph
15.5.1 Supply Chain Management using Knowledge Graph
15.5.2 DRKG: A Knowledge Graph Approach to Drug Repurposing for COVID-19
15.6 Conclusion
References
Chapter 16 : GRAPH REPRESENTATION LEARNING IN WIRELESS COMMUNICATION AND TOURIST MOVEMENT ANALYSIS
16.1 A Graph-Theoretic Framework for Analyzing Tourist Flows from Social Media Data
16.2 Knowledge-Driven Graph Learning for Next-Generation Wireless Networks
16.3 Applications of Graph Representation Learning in Wireless Networks and Urban Tourism
16.4 Other areas of GRL applications
16.5 Conclusion
References
