English

The energy landscape of a simple neural network

Machine Learning 2017-06-23 v1 Machine Learning

Abstract

We explore the energy landscape of a simple neural network. In particular, we expand upon previous work demonstrating that the empirical complexity of fitted neural networks is vastly less than a naive parameter count would suggest and that this implicit regularization is actually beneficial for generalization from fitted models.

Keywords

Cite

@article{arxiv.1706.07101,
  title  = {The energy landscape of a simple neural network},
  author = {Anthony Collins Gamst and Alden Walker},
  journal= {arXiv preprint arXiv:1706.07101},
  year   = {2017}
}

Comments

17 pages, 15 figures