Self-Bounding Majority Vote Learning Algorithms by the Direct Minimization of a Tight PAC-Bayesian C-Bound.
Paul ViallardPascal GermainAmaury HabrardEmilie MorvantPublished in: CoRR (2021)
Keyphrases
- majority vote
- pac bayesian
- upper bound
- rademacher complexity
- learning algorithm
- lower bound
- distribution free
- risk bounds
- generalization error
- generalization bounds
- vc dimension
- worst case
- data dependent
- learning problems
- classifier ensemble
- classifier combination
- sample complexity
- machine learning algorithms
- active learning
- error bounds
- fusion methods
- objective function
- machine learning
- generalization ability
- random selection
- training data
- learning tasks
- statistical learning theory
- supervised learning
- concept classes
- concept class
- sample size
- learning rate
- pac learning
- data sets
- unlabeled data
- unsupervised learning
- decision trees
- neural network