Automatic Setting of DNN Hyper-Parameters by Mixing Bayesian Optimization and Tuning Rules.
Michele FraccaroliEvelina LammaFabrizio RiguzziPublished in: LOD (1) (2020)
Keyphrases
- hyperparameters
- bayesian inference
- posterior distribution
- parameter optimization
- model selection
- cross validation
- maximum likelihood
- bayesian methods
- random sampling
- gaussian processes
- conjugate priors
- grid search
- prior information
- parameter settings
- gaussian process
- closed form
- bayesian learning
- support vector
- sample size
- incremental learning
- bayesian framework
- em algorithm
- variational bayes
- noise level
- bayesian networks
- missing values
- active learning
- parameter values
- incomplete data
- posterior probability
- genetic algorithm ga
- probabilistic model
- evolutionary algorithm