Machine Learning with Graphs
Machine Learning with Graphs
Course Description:
While existing deep learning methods have seen success with image, video, and text data, many real-world applications fall outside these domains. Graph neural networks—also known as deep learning on graphs or geometric deep learning—have emerged as one of the fastest-growing research areas in machine learning, particularly in deep learning. This course aims to provide a comprehensive introduction to graph neural networks, covering their foundations, cutting-edge developments, and applications.
Tentative Course Outline:
| Week | Content |
|---|---|
| Week 1 | - Course Introduction; Graph Basics and Machine Learning Basics - Survey 1 - Course project introduction |
| Week 2 | - Graph Convolution, Spatial and Spectral Graph Neural Networks - Quiz 1 |
| Week 3 | - Message Passing GNNs and other GNNs - HW1 release - Quiz 2 |
| Week 4 | - GNN Scientific Computing - Group member list finalize and initial project topic selection - Quiz 3 |
| Week 5 | - Graph Neural Network Node/Edge Level Tasks - HW1 due - HW2 release |
| Week 6 | - Graph Neural Network Graph Level Tasks - Mid-term project proposal presentation - Project proposal report due - Quiz 4 |
| Week 7 | - Graph Structure Learning - HW2 due - HW3 release |
| Week 8 | - Adversarial Attack and Defense |
| Week 9 | - Fast and Scalable GNNs - HW3 due |
| Week 10 | - Implicit Graph Neural Networks |
| Week 11 | - Graphs and Neural Network Architectures - Project progress check |
| Week 12 | - Theoretical Analysis |
| Week 13 | - Algorithmic Reasoning |
| Week 14 | - Graph Generative Models - Survey 2 |
| Week 15 | - Project presentation - Project report due |