Rate-adaptive model selection over a collection of black-box contextual bandit algorithms.
Aurélien F. BibautAntoine ChambazMark J. van der LaanPublished in: CoRR (2020)
Keyphrases
- black box
- model selection
- black boxes
- contextual bandit
- cross validation
- white box
- parameter estimation
- bayesian methods
- upper confidence bound
- integration testing
- mixture model
- statistical inference
- gaussian process
- selection criterion
- machine learning
- regression model
- feature selection
- learning algorithm
- meta learning
- hyperparameters
- error estimation
- recommender systems
- model selection criteria
- feature space