We propose a greedy strategy to spectrally train a deep network for multi-class classification. Each layer is defined as a composition of linear weights with the feature map of a Gaussian kernel acting as the activation function. At each layer, the linear weights are learned by maximizing the dependence between the layer output and the labels using the Hilbert Schmidt Independence Criterion (HSIC). By constraining the solution space on the Stiefel Manifold, we demonstrate how our network construct (Kernel Dependence Network or KNet) can be solved spectrally while leveraging the eigenvalues to automatically find the width and the depth of the network. We theoretically guarantee the existence of a solution for the global optimum while providing insight into our network's ability to generalize.
@article{arxiv.2011.03320,
title = {Kernel Dependence Network},
author = {Chieh Wu and Aria Masoomi and Arthur Gretton and Jennifer Dy},
journal= {arXiv preprint arXiv:2011.03320},
year = {2020}
}
Comments
arXiv admin note: substantial text overlap with arXiv:2006.08539