Homogeneous vector bundles and $G$-equivariant convolutional neural networks

Abstract

$G$-equivariant convolutional neural networks (GCNNs) is a geometric deep learning model for data defined on a homogeneous $G$-space $\mathcal{M}$. GCNNs are designed to respect the global symmetry in $\mathcal{M}$, thereby facilitating learning. In this paper, we analyze GCNNs on homogeneous spaces $\mathcal{M}=G/K$ in the case of unimodular Lie groups $G$ and compact subgroups $K\le G$. We demonstrate that homogeneous vector bundles is the natural setting for GCNNs. We also use reproducing kernel Hilbert spaces to obtain a precise criterion for expressing $G$-equivariant layers as convolutional layers. This criterion is then rephrased as a bandwidth criterion, leading to even stronger results for some groups.

23 pages

References in corpus (11)

• Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision
• EfficientNetV2: Smaller Models and Faster Training
• Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges
• High-Performance Large-Scale Image Recognition Without Normalization
• CvT: Introducing Convolutions to Vision Transformers
• Lattice gauge equivariant convolutional neural networks
• Gauge equivariant neural networks for quantum lattice gauge theories
• Machine-learning physics from unphysics: Finding deconfinement temperature in lattice Yang-Mills theories from outside the scaling window
• Theoretical Aspects of Group Equivariant Neural Networks
• A Wigner-Eckart Theorem for Group Equivariant Convolution Kernels
• Covariance in Physics and Convolutional Neural Networks

Cited by in corpus (2)

• Steerable Partial Differential Operators for Equivariant Neural Networks
• Computing equivariant matrices on homogeneous spaces for Geometric Deep Learning and Automorphic Lie Algebras