Group Equivariant Deep Learning - 2022

Dive into an 8-hour lecture series on Group Equivariant Deep Learning from the University of Amsterdam (2022). Explore fundamental concepts of group theory, including product, inverse, and representations. Learn about regular group convolutional neural networks and their historical development. Delve into advanced topics such as steerable kernels, basis functions, and irreducible representations. Discover the applications of group convolutions in neural networks and their importance in equivariant deep learning. Investigate SE(3) equivariant graph neural networks, tensor products, and message passing techniques. Gain insights into SO(3) irreps, Wigner-D matrices, and Clebsch-Gordan coefficients. Examine literature on 3D steerable graph neural networks, regular equivariant graph neural networks, and gauge equivariant graph neural networks.

Group Equivariant Deep Learning - Lecture 1.1: Introduction.
Group Equivariant Deep Learning - Lecture 1.2: Group theory (product, inverse, representations).
Group Equivariant Deep Learning - Lecture 1.3: Regular group convolutional neural networks.
Group Equivariant Deep Learning - Lecture 1.4: Example.
Group Equivariant Deep Learning - Lecture 1.5: A Brief History of G-CNNs.
Group Equivariant Deep Learning - Lecture 1.6: Group theory (Homogeneous/quotient spaces).
Group Equivariant Deep Learning - Lecture 1.7: Group convolutions are all you need.
Group Equivariant Deep Learning - Lecture 2.1: Steerable kernels/basis functions.
Group Equivariant Deep Learning - Lecture 2.2: Revisiting Regular G-Convs with Steerable Kernels.
Group Equivariant Deep Learning - Lecture 2.3: Group Theory (Irreducible representations, Fourier).
Group Equivariant Deep Learning - Lecture 2.4: Group Theory (Induced representation, feature fields).
Group Equivariant Deep Learning - Lecture 2.5: Steerable group convolutions.
Group Equivariant Deep Learning - Lecture 2.6: Activation Functions for Steerable G-CNNs.
Group Equivariant Deep Learning - Lecture 2.7: Derivation of Harmonic Networks from Regular G-Convs.
Group Equivariant Deep Learning - Lecture 3.1: Motivation for SE(3) equivariant graph NNs.
Group Equivariant Deep Learning - Lecture 3.2: Equivariant message passing as non-linear convolution.
Group Equivariant Deep Learning - Lecture 3.3: Tensor products as conditional linear layers.
Group Equivariant Deep Learning - Lecture 3.4: Group Theory (SO(3) irreps, Wigner-D, Clebsch-Gordan).
Group Equivariant Deep Learning - Lecture 3.5: Literature survey (3D Steerable graph NNs).
Group Equivariant Deep Learning - Lecture 3.6: Literature survey (Regular equivariant graph NNs).
Group Equivariant Deep Learning - Lecture 3.7: Gauge equivariant graph NNs.