ML Anthology

A PAC-Bayes Risk Bound for General Loss Functions

Pascal Germain, Alexandre Lacasse, François Laviolette, Mario Marchand
NeurIPS 2006 pp. 449-456
/neurips/2006/germain2006neurips-pacbayes/

Abstract

We provide a PAC-Bayesian bound for the expected loss of convex combinations of classifiers under a wide class of loss functions (which includes the exponential loss and the logistic loss). Our numerical experiments with Adaboost indicate that the proposed upper bound, computed on the training set, behaves very similarly as the true loss estimated on the testing set.

PDF NeurIPS Semantic Scholar

Cite

Text

Germain et al. "A PAC-Bayes Risk Bound for General Loss Functions." Neural Information Processing Systems, 2006.

Markdown

[Germain et al. "A PAC-Bayes Risk Bound for General Loss Functions." Neural Information Processing Systems, 2006.](https://mlanthology.org/neurips/2006/germain2006neurips-pacbayes/)

BibTeX

@inproceedings{germain2006neurips-pacbayes,
  title     = {{A PAC-Bayes Risk Bound for General Loss Functions}},
  author    = {Germain, Pascal and Lacasse, Alexandre and Laviolette, François and Marchand, Mario},
  booktitle = {Neural Information Processing Systems},
  year      = {2006},
  pages     = {449-456},
  url       = {https://mlanthology.org/neurips/2006/germain2006neurips-pacbayes/}
}