Unsupervised Layer-Wise Model Selection in Deep Neural Networks.
Ludovic ArnoldHélène Paugam-MoisyMichèle SebagPublished in: ECAI (2010)
Keyphrases
- model selection
- neural network
- unsupervised learning
- multi layer
- cross validation
- competitive learning
- bayesian learning
- hyperparameters
- machine learning
- parameter estimation
- regression model
- mixture model
- statistical learning
- statistical inference
- sample size
- meta learning
- error estimation
- feature selection
- pairwise
- selection criterion
- motion segmentation
- artificial neural networks
- generalization error
- gaussian process
- leave one out cross validation
- variable selection
- supervised learning
- bayesian methods
- semi supervised
- information criterion
- marginal likelihood
- parameter determination