ML Anthology

A Noninformative Prior for Neural Networks

Herbert K. H. Lee
MLJ 2003 pp. 197-212
doi:10.1023/A:1020258113913 /mlj/2003/lee2003mlj-noninformative/

Abstract

While many implementations of Bayesian neural networks use large, complex hierarchical priors, in much of modern Bayesian statistics, noninformative (flat) priors are very common. This paper introduces a noninformative prior for feed-forward neural networks, describing several theoretical and practical advantages of this approach. In particular, a simpler prior allows for a simpler Markov chain Monte Carlo algorithm. Details of MCMC implementation are included.

PDF MLJ Semantic Scholar

Cite

Text

Lee. "A Noninformative Prior for Neural Networks." Machine Learning, 2003. doi:10.1023/A:1020258113913

Markdown

[Lee. "A Noninformative Prior for Neural Networks." Machine Learning, 2003.](https://mlanthology.org/mlj/2003/lee2003mlj-noninformative/) doi:10.1023/A:1020258113913

BibTeX

@article{lee2003mlj-noninformative,
  title     = {{A Noninformative Prior for Neural Networks}},
  author    = {Lee, Herbert K. H.},
  journal   = {Machine Learning},
  year      = {2003},
  pages     = {197-212},
  doi       = {10.1023/A:1020258113913},
  volume    = {50},
  url       = {https://mlanthology.org/mlj/2003/lee2003mlj-noninformative/}
}