Clebsch-Gordan Nets: a Fully Fourier Space Spherical Convolutional Neural Network

Abstract

Recent work by Cohen \emph{et al.} has achieved state-of-the-art results for learning spherical images in a rotation invariant way by using ideas from group representation theory and noncommutative harmonic analysis. In this paper we propose a generalization of this work that generally exhibits improved performace, but from an implementation point of view is actually simpler. An unusual feature of the proposed architecture is that it uses the Clebsch--Gordan transform as its only source of nonlinearity, thus avoiding repeated forward and backward Fourier transforms. The underlying ideas of the paper generalize to constructing neural networks that are invariant to the action of other compact groups.

arXiv commentCamera ready version for the proceedings of the thirty-second conference on Neural Information Processing Systems (NIPS), Montreal, Canada, 2018

References in corpus (2)

• On the Generalization of Equivariance and Convolution in Neural Networks to the Action of Compact Groups
• Equivariance Through Parameter-Sharing

Cited by in corpus (16)

• Machine learning and the physical sciences
• On the Generalization of Equivariance and Convolution in Neural Networks to the Action of Compact Groups
• 3D Steerable CNNs: Learning Rotationally Equivariant Features in Volumetric Data
• Cormorant: Covariant Molecular Neural Networks
• Machine Learning on Neutron and X-Ray Scattering
• Informing Geometric Deep Learning with Electronic Interactions to Accelerate Quantum Chemistry
• DeepCMB: Lensing Reconstruction of the Cosmic Microwave Background with Deep Neural Networks
• Re-evaluating Evaluation
• Homogeneous vector bundles and $G$-equivariant convolutional neural networks
• Neural Star Domain as Primitive Representation
• Rotation-Invariant Autoencoders for Signals on Spheres
• VolterraNet: A higher order convolutional network with group equivariance for homogeneous manifolds
• Equivariant Networks for Pixelized Spheres
• Discriminative Learning of Similarity and Group Equivariant Representations
• Physics-Augmented Learning: A New Paradigm Beyond Physics-Informed Learning
• On Universalized Adversarial and Invariant Perturbations