Section outline

• Select section Day 1: Introduction to Graph Neural Networks

  Day 1: Introduction to Graph Neural Networks

  Collapse all Expand all

  Session 1: Graph Theory & Deep Learning (3.5 hours)

  • Basic Graph Theory: Nodes, Edges, Directed & Undirected Graphs
  • Graph Representation: Adjacency Matrix, Adjacency List, Edge List
  • Common Graph Algorithms: DFS, BFS, Dijkstra’s Algorithm
  • Neural Network Basics & Why GNNs for Graph Data

  Hands-on Exercise: DFS & BFS on Graph Data

  Session 2: Foundations of Graph Neural Networks (4.5 hours)

  • What is a Graph Neural Network (GNN)?
  • Graph Representation Learning & GCNs
  • Graph Message Passing & Aggregation Functions
  • Node Embedding & Graph Embedding

  Hands-on Exercise: Implementing Graph Convolution Networks (GCNs)
• Select section Day 2: Advanced GNN Architectures & Applications

  Day 2: Advanced GNN Architectures & Applications

  Session 1: GNN Architectures & Graph Attention Networks (3.5 hours)

  • Graph Convolutional Networks (GCNs) & Spectral Methods
  • Graph Attention Networks (GATs) & Multi-Head Attention
  • Neighborhood Sampling Techniques

  Hands-on Exercise: Implementing GCNs & GATs

  Session 2: Advanced GNN Topics & Deployment (4.5 hours)

  • Graph Autoencoders & Variational Graph Autoencoders (VGAE)
  • Relational Graph Convolutional Networks (R-GCNs)
  • Graph Isomorphism Networks (GINs)
  • Scalability Issues & Dynamic Graphs
  • Semi-supervised & Self-supervised Learning for GNNs

  Hands-on Exercise: Building Explainable GNN Models