Corpus ID: 49406331

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

@inproceedings{Kondor2018ClebschGordanNA,
  title={Clebsch-Gordan Nets: a Fully Fourier Space Spherical Convolutional Neural Network},
  author={Risi Kondor and Z. Lin and S. Trivedi},
  booktitle={NeurIPS},
  year={2018}
}
  • Risi Kondor, Z. Lin, S. Trivedi
  • Published in NeurIPS 2018
  • Mathematics, Computer Science
  • 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. [...] Key Method 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…Expand Abstract
    On the Generalization of Equivariance and Convolution in Neural Networks to the Action of Compact Groups
    132
    Deep Scale-spaces: Equivariance Over Scale
    27
    Theoretical Aspects of Group Equivariant Neural Networks
    4
    SO(2)-equivariance in Neural networks using tensor nonlinearity
    2
    Spin-Weighted Spherical CNNs
    Learning to Convolve: A Generalized Weight-Tying Approach
    5
    A General Theory of Equivariant CNNs on Homogeneous Spaces
    53
    Gauge Equivariant Spherical CNNs
    Gauge Equivariant Convolutional Networks and the Icosahedral CNN
    69

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 41 REFERENCES
    On the Generalization of Equivariance and Convolution in Neural Networks to the Action of Compact Groups
    132
    Group Equivariant Convolutional Networks
    483
    Geodesic Convolutional Neural Networks on Riemannian Manifolds
    411
    Steerable CNNs
    135
    Equivariance Through Parameter-Sharing
    68
    Deep Symmetry Networks
    166
    Spherical CNNs
    204
    Learning SO(3) Equivariant Representations with Spherical CNNs
    70
    Polar Transformer Networks
    66