Corpus ID: 29153296

Butterfly-Net: Optimal Function Representation Based on Convolutional Neural Networks

@article{Li2018ButterflyNetOF,
  title={Butterfly-Net: Optimal Function Representation Based on Convolutional Neural Networks},
  author={Yingzhou Li and X. Cheng and Jianfeng Lu},
  journal={ArXiv},
  year={2018},
  volume={abs/1805.07451}
}
  • Yingzhou Li, X. Cheng, Jianfeng Lu
  • Published 2018
  • Computer Science, Mathematics
  • ArXiv
  • Deep networks, especially Convolutional Neural Networks (CNNs), have been successfully applied in various areas of machine learning as well as to challenging problems in other scientific and engineering fields. This paper introduces Butterfly-net, a low-complexity CNN with structured and sparse across-channel connections, which aims at an optimal hierarchical function representation of the input signal. Theoretical analysis of the approximation power of Butterfly-net to the Fourier… CONTINUE READING

    Figures, Tables, and Topics from this paper.

    Butterfly-Net2: Simplified Butterfly-Net and Fourier Transform Initialization
    SwitchNet: a neural network model for forward and inverse scattering problems
    23
    BCR-Net: a neural network based on the nonstandard wavelet form
    14
    A multiscale neural network based on hierarchical matrices
    28
    VerSe: A Vertebrae Labelling and Segmentation Benchmark
    5

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 50 REFERENCES
    Universality of Deep Convolutional Neural Networks
    54
    Optimal Approximation with Sparsely Connected Deep Neural Networks
    107
    On the Number of Linear Regions of Deep Neural Networks
    622
    PDE-Net: Learning PDEs from Data
    193
    On the Expressive Power of Deep Learning: A Tensor Analysis
    210
    A multiscale neural network based on hierarchical nested bases
    17
    U-Net: Convolutional Networks for Biomedical Image Segmentation
    14804
    Why Deep Neural Networks for Function Approximation?
    160
    Deep vs. shallow networks : An approximation theory perspective
    185
    The Power of Depth for Feedforward Neural Networks
    370