|
| 1 | +title: 'Tutorial 6: Basics of Graph Neural Networks' |
| 2 | +author: Phillip Lippe |
| 3 | +created: 2021-06-07 |
| 4 | +updated: 2023-03-14 |
| 5 | +license: CC BY-SA |
| 6 | +build: 0 |
| 7 | +tags: |
| 8 | +- Graph |
| 9 | +description: 'In this tutorial, we will discuss the application of neural networks |
| 10 | + on graphs. |
| 11 | +
|
| 12 | + Graph Neural Networks (GNNs) have recently gained increasing popularity in both |
| 13 | + applications and research, |
| 14 | +
|
| 15 | + including domains such as social networks, knowledge graphs, recommender systems, |
| 16 | + and bioinformatics. |
| 17 | +
|
| 18 | + While the theory and math behind GNNs might first seem complicated, |
| 19 | +
|
| 20 | + the implementation of those models is quite simple and helps in understanding the |
| 21 | + methodology. |
| 22 | +
|
| 23 | + Therefore, we will discuss the implementation of basic network layers of a GNN, |
| 24 | +
|
| 25 | + namely graph convolutions, and attention layers. |
| 26 | +
|
| 27 | + Finally, we will apply a GNN on semi-supervised node classification and molecule |
| 28 | + categorization. |
| 29 | +
|
| 30 | + This notebook is part of a lecture series on Deep Learning at the University of |
| 31 | + Amsterdam. |
| 32 | +
|
| 33 | + The full list of tutorials can be found at https://uvadlc-notebooks.rtfd.io. |
| 34 | +
|
| 35 | + ' |
| 36 | +requirements: |
| 37 | +- torch-scatter |
| 38 | +- torch-sparse |
| 39 | +- torch-cluster |
| 40 | +- torch-spline-conv |
| 41 | +- torch-geometric>=2.0.0,<2.5.0 |
| 42 | +- lightning>=2.0.0 |
| 43 | +pip__find-link: |
| 44 | +- https://pytorch-geometric.com/whl/torch-%(TORCH_MAJOR_DOT_MINOR)s.0+%(DEVICE)s.html |
| 45 | +accelerator: |
| 46 | +- GPU |
| 47 | +environment: |
| 48 | +- matplotlib==3.8.4 |
| 49 | +- numpy==1.26.4 |
| 50 | +- torch-sparse==0.6.18+pt20cu118 |
| 51 | +- lightning==2.3.3 |
| 52 | +- urllib3==2.2.2 |
| 53 | +- torchmetrics==1.2.1 |
| 54 | +- torch-scatter==2.1.2+pt20cu118 |
| 55 | +- torch-cluster==1.6.3+pt20cu118 |
| 56 | +- torch_geometric==2.4.0 |
| 57 | +- torch==2.0.1+cu118 |
| 58 | +- torch-spline-conv==1.2.2+pt20cu118 |
| 59 | +- setuptools==69.0.3 |
| 60 | +- pytorch-lightning==1.5.3 |
| 61 | +- ipython==8.16.1 |
| 62 | +published: '2024-07-22T15:49:36.460618' |
0 commit comments