Applied Deep Learning 2025 - Lecture 11 - Graph Neural Networks

Combining machine learning with graphs opens up an entirely new world of possibilities. In this lecture we're diving into that world and explore how you can formulate problems to make use of graph neural networks, which mechanisms are used to train such models, and which kind of problems you might be able to solve with them.

Complete Playlist:    • Applied Deep Learning 2025 - Lecture 0 - P...  

00:00 - Start
00:47 - Graphs as inputs for Neural Networks
02:03 - What are Graph Neural Networks?
03:56 - Task 1: Node Classification
04:49 - Task 2: Graph Classification
05:54 - Task 3: Link Classification
06:54 - How to get the input graph?
12:01 - Getting a "better" Graph
16:31 - Convolutional Graph Neural Networks
22:56 - The CORA dataset
24:28 - GNNs in Code
33:20 - Graph Attentional Networks
35:50 - Neural Message Passing
39:31 - Frameworks
43:07 - Applications
46:43 - Summary

== Literature ==
1. Menzli, Graph Neural Networks: Libraries, Tools, and Learning Resources. 2021
2. Hamilton, Graph Representation Learning. 2020
3. Stokes et al., A Deep Learning Approach to Antibiotic Discovery. 2020
4. Petar Veličković, Intro to graph neural networks. 2021
5. Petar Veličković, Theoretical Foundations of Graph Neural Networks. 2021
6. Petar Veličković et al., Graph Attention Networks. 2018
7. Kipf et al., Semi-Supervised Classification with Graph Convolutional Networks. 2017
8. Gordić, Graph Attention Networks in PyTorch (pytorch-GAT). 2020
9. Allamanis, An Introduction to Graph Neural Networks: Models and Applications. 2019
10. Joshi, Transformers are Graph Neural Networks. 2020
11. Barragao, CORA.
12. Kipf, Graph Convolutional Networks. 2016
13. Neys, Intro to DeepMind’s GraphNet. 2021
14. https://graphdeeplearning.github.io/
15. http://web.stanford.edu/class/cs224w/
16. Jumper et al. Highly accurate protein structure prediction with AlphaFold, 2021