ML Anthology

A PAC-Bayes Sample-Compression Approach to Kernel Methods

Pascal Germain, Alexandre Lacoste, François Laviolette, Mario Marchand, Sara Shanian
ICML 2011 pp. 297-304
/icml/2011/germain2011icml-pac/

Abstract

We propose a PAC-Bayes sample compression approach to kernel methods that can accommodate any bounded similarity function and show that the support vector machine (SVM) classifier is a particular case of a more general class of data-dependent classifiers known as majority votes of sample-compressed classifiers. We provide novel risk bounds for these majority votes and learning algorithms that minimize these bounds.

PDF Semantic Scholar

Cite

Text

Germain et al. "A PAC-Bayes Sample-Compression Approach to Kernel Methods." International Conference on Machine Learning, 2011.

Markdown

[Germain et al. "A PAC-Bayes Sample-Compression Approach to Kernel Methods." International Conference on Machine Learning, 2011.](https://mlanthology.org/icml/2011/germain2011icml-pac/)

BibTeX

@inproceedings{germain2011icml-pac,
  title     = {{A PAC-Bayes Sample-Compression Approach to Kernel Methods}},
  author    = {Germain, Pascal and Lacoste, Alexandre and Laviolette, François and Marchand, Mario and Shanian, Sara},
  booktitle = {International Conference on Machine Learning},
  year      = {2011},
  pages     = {297-304},
  url       = {https://mlanthology.org/icml/2011/germain2011icml-pac/}
}