Information-Theoretic Bounds on the Moments of the Generalization Error of Learning Algorithms.
Gholamali AminianLaura ToniMiguel R. D. RodriguesPublished in: ISIT (2021)
Keyphrases
- information theoretic
- generalization error
- learning algorithm
- upper bound
- algorithmic stability
- generalization error bounds
- training error
- uniform convergence
- learning machines
- sample complexity
- information theory
- mutual information
- active learning
- cross validation
- lower bound
- linear classifiers
- binary classification
- model selection
- target function
- vc dimension
- training data
- sample size
- supervised learning
- information bottleneck
- worst case
- training set
- learning rate
- information theoretic measures
- computational learning theory
- learning problems
- learning tasks
- generalization bounds
- machine learning algorithms
- machine learning
- generalization ability
- learning process
- data sets
- reinforcement learning
- similarity measure
- back propagation
- bregman divergences
- training samples
- semi supervised learning
- labeled data