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