فهرست مطالب

• Graph Neural Networks in Action
• copyright
• contents
• dedication
• foreword
• preface
• acknowledgments
• about this book
• about the authors
• about the cover illustration
• Part 1 First steps
• 1 Discovering graph neural networks
• 1.1 Goals of this book
• 1.1.1 Catching up on graph fundamentals
• 1.2 Graph-based learning
• 1.2.1 What are graphs?
• 1.2.2 Different types of graphs
• 1.2.3 Graph-based learning
• 1.2.4 What is a GNN?
• 1.2.5 Differences between tabular and graph data
• 1.3 GNN applications: Case studies
• 1.3.1 Recommendation engines
• 1.3.2 Drug discovery and molecular science
• 1.3.3 Mechanical reasoning
• 1.4 When to use a GNN?
• 1.4.1 Implicit relationships and interdependencies
• 1.4.2 High dimensionality and sparsity
• 1.4.3 Complex, nonlocal interactions
• 1.5 Understanding how GNNs operate
• 1.5.1 Mental model for training a GNN
• 1.5.2 Unique mechanisms of a GNN model
• 1.5.3 Message passing
• 2 Graph embeddings
• 2.1 Creating embeddings with Node2Vec
• 2.1.1 Loading data, setting parameters, and creating embeddings
• 2.1.2 Demystifying embeddings
• 2.1.3 Transforming and visualizing the embeddings
• 2.1.4 Beyond visualization: Applications and considerations of N2V embeddings
• 2.2 Creating embeddings with a GNN
• 2.2.1 Constructing the embeddings
• 2.2.2 GNN vs. N2V embeddings
• 2.3 Using node embeddings
• 2.3.1 Data preprocessing
• 2.3.2 Random forest classification
• 2.3.3 Embeddings in an end-to-end model
• 2.4 Under the Hood
• 2.4.1 Representations and embeddings
• 2.4.2 Transductive and inductive methods
• 2.4.3 N2V: Random walks across graphs
• 2.4.4 Message passing as deep learning
• Part 2 Graph neural networks
• 3 Graph convolutional networks and GraphSAGE
• 3.1 Predicting consumer product categories
• 3.1.1 Loading and processing the data
• 3.1.2 Creating our model classes
• 3.1.3 Model training
• 3.1.4 Model performance analysis
• 3.1.5 Our first product bundle
• 3.2 Aggregation methods
• 3.2.1 Neighborhood aggregation
• 3.2.2 Advanced aggregation tools
• 3.2.3 Practical considerations in applying aggregation
• 3.3 Further optimizations and refinements
• 3.3.1 Dropout
• 3.3.2 Model depth
• 3.3.3 Improving the baseline model’s performance
• 3.3.4 Revisiting the Marcelina product bundle
• 3.4 Under the hood
• 3.4.1 Convolution methods
• 3.4.2 Message passing
• 3.4.3 GCN aggregation function
• 3.4.4 GCN in PyTorch Geometric
• 3.4.5 Spectral vs. spatial convolution
• 3.4.6 GraphSAGE aggregation function
• 3.4.7 GraphSAGE in PyTorch Geometric
• 3.5 Amazon Products dataset
• 4 Graph attention networks
• 4.1 Detecting spam and fraudulent reviews
• 4.2 Exploring the review spam dataset
• 4.2.1 Explaining the node features
• 4.2.2 Exploratory data analysis
• 4.2.3 Exploring the graph structure
• 4.2.4 Exploring the node features
• 4.3 Training baseline models
• 4.3.1 Non-GNN baselines
• 4.3.2 GCN baseline
• 4.4 Training GAT models
• 4.4.1 Neighborhood loader and GAT models
• 4.4.2 Addressing class imbalance in model performance
• 4.4.3 Deciding between GAT and XGBoost
• 4.5 Under the hood
• 4.5.1 Explaining attention and GAT models
• 4.5.2 Over-smoothing
• 4.5.3 Overview of key GAT equations
• 5 Graph autoencoders
• 5.1 Generative models: Learning how to generate
• 5.1.1 Generative and discriminative models
• 5.1.2 Synthetic data
• 5.2 Graph autoencoders for link prediction
• 5.2.1 Review of the Amazon Products dataset from chapter 3
• 5.2.2 Defining a graph autoencoder
• 5.2.3 Training a graph autoencoder to perform link prediction
• 5.3 Variational graph autoencoders
• 5.3.1 Building a variational graph autoencoder
• 5.3.2 When to use a variational graph autoencoder
• 5.4 Generating graphs using GNNs
• 5.4.1 Molecular graphs
• 5.4.2 Identifying new drug candidates
• 5.4.3 VGAEs for generating graphs
• 5.4.4 Generating molecules using a GNN
• 5.5 Under the hood
• 5.5.1 Understanding link prediction tasks
• 5.5.2 The inner product decoder
• Part 3 Advanced topics
• 6 Dynamic graphs: Spatiotemporal GNNs
• 6.1 Temporal models: Relations through time
• 6.2 Problem definition: Pose estimation
• 6.2.1 Setting up the problem
• 6.2.2 Building models with memory
• 6.3 Dynamic graph neural networks
• 6.3.1 Graph attention network for dynamic graphs
• 6.4 Neural relational inference
• 6.4.1 Encoding pose data
• 6.4.2 Decoding pose data using a GRU
• 6.4.3 Training the NRI model
• 6.5 Under the hood
• 6.5.1 Recurrent neural networks
• 6.5.2 Temporal adjacency matrices
• 6.5.3 Combining autoencoders with RNNs
• 6.5.4 Gumbel-Softmax
• 7 Learning and inference at scale
• 7.1 Examples in this chapter
• 7.1.1 Amazon Products dataset
• 7.1.2 GeoGrid
• 7.2 Framing problems of scale
• 7.2.1 Root causes
• 7.2.2 Symptoms
• 7.2.3 Crucial metrics
• 7.3 Techniques for tackling problems of scale
• 7.3.1 Seven techniques
• 7.3.2 General Steps
• 7.4 Choice of hardware configuration
• 7.4.1 Types of hardware choices
• 7.4.2 Choice of processor and memory size
• 7.5 Choice of data representation
• 7.6 Choice of GNN algorithm
• 7.6.1 Time and space complexity
• 7.7 Batching using a sampling method
• 7.7.1 Two concepts: Mini-batching and sampling
• 7.7.2 A glance at notable PyG samplers
• 7.8 Parallel and distributed processing
• 7.8.1 Using distributed data parallel
• 7.8.2 Code example for DDP
• 7.9 Training with remote storage
• 7.9.1 Example
• 7.10 Graph coarsening
• 7.10.1 Example
• 8 Considerations for GNN projects
• 8.1 Data preparation and project planning
• 8.1.1 Project definition
• 8.1.2 Project objectives and scope
• 8.2 Designing graph models
• 8.2.1 Get familiar with the domain and use case
• 8.2.2 Constructing the graph dataset and schemas
• 8.2.3 Creating instance models
• 8.2.4 Testing and refactoring
• 8.3 Data pipeline example
• 8.3.1 Raw data
• 8.3.2 The ETL step
• 8.3.3 Data exploration and visualization
• 8.3.4 Preprocessing and loading data into PyG
• 8.4 Where to find graph data
• Appendix A Discovering graphs
• A.1 Graph fundamentals
• A.1.1 Graph properties
• A.1.2 Characteristics of nodes and edges
• A.1.3 Categories of graphs
• A.2 Graph representations
• A.2.1 Basic graph data structures
• A.2.2 Relational databases
• A.2.3 How graphs are exposed
• A.3 Graph systems
• A.3.1 Graph databases
• A.3.2 Graph compute engines (or graph frameworks)
• A.3.3 Visualization libraries
• A.3.4 GNN libraries
• A.4 Graph algorithms
• A.4.1 Traversal and search algorithms
• A.4.2 Shortest path
• A.5 How to read GNN literature
• A.5.1 Common graph notations
• Appendix B Installing and configuring PyTorch Geometric
• B.1 Installing PyTorch Geometric
• B.1.1 On Windows/Linux
• B.1.2 On MacOS
• B.1.3 Compatibility issues
• further reading
• references