Geometric Deep Learning
Neural architectures that encode symmetry, geometry, equivariance, and three-dimensional structure.
- Understanding the spherical equivariant layers that power modern molecular neural networks, from group theory foundations to Clebsch-Gordan tensor products.