ppt课件-p a c- bayes risk boundsfor sample- compressed gibbs classifiers（p c -贝叶斯风险boundsfor样本压缩吉布斯分类器）.pptVIP

PAC-Bayes Risk Bounds for Sample-Compressed Gibbs Classifiers ICML 2005 Fran?ois Laviolette and Mario Marchand Université Laval PLAN The “traditional” PAC-Bayes theorem (for the usual data-independent setting ) The “generalized” PAC-Bayes theorem  (for the more general sample compression setting) Implications and follow-ups A result from folklore : In particular, for Gibbs classifiers: The “traditional” PAC-Bayes Theorem The Gibbs and the majority vote We have a bound for GQ but we normally use instead the Bayes classifier BQ (which is the Q-weighted majority vote classifier)