DSV: An Alignment Validation Loss for Self-supervised Outlier Model Selection.
Jaemin YooYue ZhaoLingxiao ZhaoLeman AkogluPublished in: ECML/PKDD (1) (2023)
Keyphrases
- model selection
- cross validation
- hyperparameters
- bayesian learning
- sample size
- outlier detection
- regression model
- parameter estimation
- machine learning
- model selection criteria
- error estimation
- gaussian process
- mixture model
- statistical learning
- generalization error
- selection criterion
- leave one out cross validation
- meta learning
- marginal likelihood
- feature selection
- variable selection
- parameter determination
- automatic model selection
- generalization bounds
- statistical inference
- information criterion
- active learning