Corpus ID: 4601550

Identification of Shallow Neural Networks by Fewest Samples

@article{Fornasier2018IdentificationOS,
  title={Identification of Shallow Neural Networks by Fewest Samples},
  author={M. Fornasier and J. Vyb{\'i}ral and I. Daubechies},
  journal={ArXiv},
  year={2018},
  volume={abs/1804.01592}
}
  • M. Fornasier, J. Vybíral, I. Daubechies
  • Published 2018
  • Mathematics, Computer Science
  • ArXiv
  • We address the uniform approximation of sums of ridge functions $\sum_{i=1}^m g_i(a_i\cdot x)$ on ${\mathbb R}^d$, representing the shallowest form of feed-forward neural network, from a small number of query samples, under mild smoothness assumptions on the functions $g_i$'s and near-orthogonality of the ridge directions $a_i$'s. The sample points are randomly generated and are universal, in the sense that the sampled queries on those points will allow the proposed recovery algorithms to… CONTINUE READING
    6 Citations
    Affine Symmetries and Neural Network Identifiability
    • 1
    Estimating multi-index models with response-conditional least squares
    • PDF
    Affine symmetries and neural network identifiability
    • 1
    • PDF
    Subspace power method for symmetric tensor decomposition and generalized PCA
    • 3
    • PDF

    References

    SHOWING 1-10 OF 68 REFERENCES
    Approximation by ridge functions and neural networks
    • 99
    • PDF
    Entropy and Sampling Numbers of Classes of Ridge Functions
    • 20
    • PDF
    Breaking the Curse of Dimensionality with Convex Neural Networks
    • 292
    • Highly Influential
    • PDF
    Capturing Ridge Functions in High Dimensions from Point Queries
    • 53
    • PDF
    Learning Functions of Few Arbitrary Linear Parameters in High Dimensions
    • 69
    • PDF
    Greedy Layer-Wise Training of Deep Networks
    • 1,507
    Ridgelets: estimating with ridge functions
    • 70
    • PDF
    A Fast Learning Algorithm for Deep Belief Nets
    • 11,500
    • PDF
    On the Connection Between Learning Two-Layers Neural Networks and Tensor Decomposition
    • 35
    • PDF