A PAC-Bayes Risk Bound for General Loss Functions.
Pascal GermainAlexandre LacasseFrançois LavioletteMario MarchandPublished in: NIPS (2006)
Keyphrases
- bayes risk
- general loss functions
- sample complexity
- upper and lower bounds
- upper bound
- lower bound
- theoretical analysis
- worst case
- learning problems
- generalization error
- lower and upper bounds
- special case
- sample size
- learning algorithm
- supervised learning
- loss function
- active learning
- maximum entropy
- reproducing kernel hilbert space
- optimal solution
- gaussian kernels