English

Avoiding pathologies in very deep networks

Machine Learning 2016-07-12 v3 Machine Learning

Abstract

Choosing appropriate architectures and regularization strategies for deep networks is crucial to good predictive performance. To shed light on this problem, we analyze the analogous problem of constructing useful priors on compositions of functions. Specifically, we study the deep Gaussian process, a type of infinitely-wide, deep neural network. We show that in standard architectures, the representational capacity of the network tends to capture fewer degrees of freedom as the number of layers increases, retaining only a single degree of freedom in the limit. We propose an alternate network architecture which does not suffer from this pathology. We also examine deep covariance functions, obtained by composing infinitely many feature transforms. Lastly, we characterize the class of models obtained by performing dropout on Gaussian processes.

Keywords

Cite

@article{arxiv.1402.5836,
  title  = {Avoiding pathologies in very deep networks},
  author = {David Duvenaud and Oren Rippel and Ryan P. Adams and Zoubin Ghahramani},
  journal= {arXiv preprint arXiv:1402.5836},
  year   = {2016}
}

Comments

Fixed a typo regarding number of layers

R2 v1 2026-06-22T03:14:28.006Z