Preface
Acknowledgments
Introduction to Network Embedding
The Basics of Network Embedding
Background
The Rising of Network Embedding
The Evaluation of Network Embedding
Node Classification
Link Prediction
Node Clustering
Network Embedding for General Graphs
Representative Methods
Early Work (Early 2000s โ Early 2010s)
Modern Work (After 2014)
Theory: A Unified Network Embedding Framework
K-Order Proximity
Network Embedding Framework
Observations
Method: Network Embedding Update (NEU)
Problem Formalization
Approximation Algorithm
Empirical Analysis
Datasets
Baselines and Experimental Settings
Multi-Label Classification
Link Prediction
Experimental Results Analysis
Further Reading
Network Embedding with Additional Information
Network Embedding for Graphs with Node Attributes
Overview
Method: Text-Associated DeepWalk
Low-Rank Matrix Factorization
Text-Associated DeepWalk (TADW)
Complexity Analysis
Empirical Analysis
Datasets
TADW Settings
Baseline Methods
Classifiers and Experiment Setup
Experimental Results and Analysis
Case Study
Further Reading
Revisiting Attributed Network Embedding: A GCN-Based Perspective
GCN-Based Network Embedding
Graph Convolutional Network (GCN)
Attributed Network Embedding based on GCN
Discussions
Method: Adaptive Graph Encoder
Problem Formalization
Overall Framework
Laplacian Smoothing Filter
Adaptive Encoder
Empirical Analysis
Datasets
Baseline Methods
Evaluation Metrics and Parameter Settings
Node Clustering Results
Link Prediction Results
GAE vs. LS+RA
Ablation Study
Selection of k
Visualization
Further Reading
Network Embedding for Graphs with Node Contents
Overview
Method: Context-Aware Network Embedding
Problem Formalization
Overall Framework
Structure-Based Objective
Text-Based Objective
Context-Free Text Embedding
Context-Aware Text Embedding
Optimization of CANE
Empirical Analysis
Datasets
Baselines
Evaluation Metrics and Experimental Settings
Link Prediction
Node Classification
Case Study
Further Reading
Network Embedding for Graphs with Node Labels
Overview
Method: Max-Margin DeepWalk
Problem Formalization
DeepWalk Based on Matrix Factorization
Max-Margin DeepWalk
Optimization of MMDW
Empirical Analysis
Datasets and Experiment Settings
Baseline Methods
Experimental Results and Analysis
Visualization
Further Reading
Network Embedding with Different Characteristics
Network Embedding for Community-Structured Graphs
Overview
Method: Community-Enhanced NRL
Problem Formalization
DeepWalk
Community-Enhanced DeepWalk
Empirical Analysis
Datasets
Baseline Methods
Parameter Settings and Evaluation Metrics
Node Classification
Link Prediction
Community Detection
Visualizations of Detected Communities
Further Reading
Network Embedding for Large-Scale Graphs
Overview
Method: COmpresSIve Network Embedding (COSINE)
Problem Definition
Graph Partitioning
Group Mapping
Group Aggregation
Objective Function and Optimization
Empirical Analysis
Datasets
Baselines and Experimental Settings
Link Prediction
Multi-Label Classification
Scalability
Time Efficiency
Different Partitioning Algorithms
Further Reading
Network Embedding for Heterogeneous Graphs
Overview
Method: Relation Structure-Aware HIN Embedding
Problem Formalization
Data Observations
Basic Idea
Different Models for ARs and IRs
A Unified Model for HIN Embedding
Empirical Analysis
Datasets
Baseline Methods
Parameter Settings
Node Clustering
Link Prediction
Multi-Class Classification
Comparison of Variant Models
Visualization
Further Reading
Network Embedding Applications
Network Embedding for Social Relation Extraction
Overview
Method: TransNet
Problem Formalization
Translation Mechanism
Edge Representation Construction
Overall Architecture
Prediction
Empirical Analysis
Datasets
Baselines
Evaluation Metrics and Parameter Settings
Results and Analysis
Comparison on Labels
Case Study
Further Reading
Network Embedding for Recommendation Systems on LBSNs
Overview
Method: Joint Network and Trajectory Model (JNTM)
Problem Formalization
Modeling the Construction of the Social Network
Modeling the Generation of the Mobile Trajectories
The Joint Model
Parameter Learning
Empirical Analysis
Data Collection
Evaluation Tasks and Baselines
Experimental Results on Next-Location Recommendation
Experimental Results on Friend Recommendation
Further Reading
Network Embedding for Information Diffusion Prediction
Overview
Method: Neural Diffusion Model (NDM)
Problem Formalization
Model Assumptions
Extracting Active Users with Attention Mechanism
Unifying Active User Embeddings for Prediction with Convolutional Network
Overall Architecture, Model Details, and Learning Algorithms
Empirical Analysis
Datasets
Baselines
Hyperparameter Settings
Microscopic Diffusion Prediction
Benefits from Network Embeddings
Interpretability
Further Reading
Outlook for Network Embedding
Future Directions of Network Embedding
Network Embedding Based on Advanced Techniques
Network Embedding in More Fine-Grained Scenarios
Network Embedding with Better Interpretability
Network Embedding for Applications
Bibliography
Authors' Biographies
Blank Page