GDL Course

As part of the African Master’s in Machine Intelligence (AMMI), we have delivered a course on Geometric Deep Learing (GDL100), which closely follows the contents of our GDL proto-book. We make all materials and artefacts from this course publicly available, as companion material for our proto-book, as well as a way to dive deeper into some of the contents for future iterations of the book.

This course was also delivered, with all materials available, in 2021.

All lecture recordings

Lecture 1: Introduction Michael M. Bronstein Recording Slides

Lecture 2: High-Dimensional Learning Joan Bruna Recording Slides

Lecture 3: Geometric Priors I Taco Cohen Recording Slides

Lecture 4: Geometric Priors II Joan Bruna Recording Slides

Lecture 5: Graphs & Sets I Petar Veličković Recording Slides

Lecture 6: Graphs & Sets II Petar Veličković Recording Slides

Lecture 7: Grids Joan Bruna Recording Slides

Lecture 8: Groups Taco Cohen Recording Slides

Lecture 9: Geodesics & Manifolds Michael M. Bronstein Recording Slides

Lecture 10: Gauges Taco Cohen Recording Slides

Lecture 11: Beyond Groups Petar Veličković Recording Slides

Lecture 12: Conclusions Michael M. Bronstein Recording Slides

Tutorial 1: Introduction to (Expressive) GNNs Cristian Bodnar, Iulia Duță, Paul Scherer Colab

Tutorial 2: Group Equivariant Neural Networks Gabriele Cesa Colab

Tutorial 3: Geometric GNNs Charlie Harris, Chaitanya Joshi, Ramon Viñas Colab

Seminar 1: Graph neural networks through the lens of multi-particle dynamics and gradient flows Francesco Di Giovanni Recording Slides

Seminar 2: Subgraphs for more expressive GNNs Fabrizio Frasca Recording Slides

Seminar 3: Equivariance in Machine Learning Geordie Williamson Recording Slides

Seminar 4: Neural Sheaf Diffusion: A Topological Perspective on Heterophily and Oversmoothing in GNNs Cristian Bodnar Recording Slides

Seminar 5: Highly accurate protein structure prediction with AlphaFold Russ Bates Recording Slides