Strengthened Information-theoretic Bounds on the Generalization Error.
Ibrahim IssaAmedeo Roberto EspositoMichael GastparPublished in: CoRR (2019)
Keyphrases
- information theoretic
- generalization error
- upper bound
- generalization error bounds
- training error
- information theory
- mutual information
- lower bound
- uniform convergence
- binary classification
- sample complexity
- model selection
- classification error
- cross validation
- learning machines
- learning algorithm
- active learning
- vc dimension
- training data
- theoretic framework
- linear classifiers
- sample size
- generalization bounds
- jensen shannon divergence
- worst case
- training set
- information bottleneck
- information theoretic measures
- target function
- supervised learning
- bregman divergences
- image processing
- boosting algorithms
- error bounds
- training samples
- pairwise
- feature extraction
- decision trees
- feature selection