Course - Geometric Deep Learning HT26

• Course webpage
• Lecture notes
  • Course schedule
  • 1, Introduction
  • 2, ML in high dimensions
• Lecture notes (old)
  • 1, Introduction
  • 2, ML in high dimensions
  • 3+4, Geometric priors
  • 5, Sets
  • 6+7, Graphs
  • 8, Graph foundation models (old)
  • 9+10, Grids (old)
  • 11, Groups and homogenous spaces (old)
  • 12, Geometric graphs and geometric GNNs (old)
  • 13, Manifolds (old)
  • 14, Gauges (old)
  • 15, Spectral methods (old)
  • 16, Physics-inspired methods (old)
• Lecture recordings
  • 2024-2025
  • 2025-2026
• Other resources:
  • Mathematical Background for GDL
  • Equivariant and coordinate independent convolutional networks
• Other courses this term: Courses HT26^U
• Related courses:
  • Course - Machine Learning MT23^U
  • Course - Uncertainty in Deep Learning MT25^U
  • Course - Theories of Deep Learning MT25^U
  • Course - Optimisation for Data Science HT25^U

Notes

• Notes - Geometric Deep Learning HT26, Core idea^U
• Notes - Geometric Deep Learning HT26, Mathematical background^U
• Notes - Geometric Deep Learning HT26, Invariance and equivariance^U
• Notes - Geometric Deep Learning HT26, Scale separation^U
• Notes - Geometric Deep Learning HT26, Sets^U
• Notes - Geometric Deep Learning HT26, Graphs^U

Practicals

• 1, Invariance
• 2, Graph foundation models (old)
• 3, Spectral transformers (old)
• 4, Invariant and equivariant GNNs (old)
• 5, Neural diffusion (old)

Problem Sheets

• Sheet for lectures 3 & 4, redacted^?
• Sheet for lectures 9 & 10, redacted^?

To-Do List