Hot Swapping for Online Adaptation of Optimization Hyperparameters.
Kevin BacheDennis DeCostePadhraic SmythPublished in: ICLR (Workshop) (2015)
Keyphrases
- hyperparameters
- model selection
- grid search
- cross validation
- closed form
- bayesian framework
- prior information
- gaussian process
- bayesian inference
- support vector
- parameter optimization
- incremental learning
- random sampling
- em algorithm
- incomplete data
- sample size
- maximum likelihood
- gaussian processes
- regularization parameter
- generalization ability
- noise level
- maximum a posteriori
- parameter settings
- data sets
- learning process
- prior knowledge
- denoising
- multiscale