Taming Hyper-parameters in Deep Learning Systems.
Luo MaiAlexandros KoliousisGuo LiAndrei-Octavian BrabetePeter R. PietzuchPublished in: ACM SIGOPS Oper. Syst. Rev. (2019)
Keyphrases
- learning systems
- hyperparameters
- model selection
- cross validation
- closed form
- bayesian inference
- bayesian framework
- support vector
- random sampling
- prior information
- sample size
- posterior distribution
- noise level
- maximum likelihood
- em algorithm
- incremental learning
- variational bayes
- machine learning
- incomplete data
- learning environment
- computer supported
- maximum a posteriori
- bayesian methods
- learning styles
- learning process
- missing values
- technology enhanced learning systems
- adaptive systems
- machine learning systems
- statistical machine learning
- training set
- active learning
- data sets
- conjugate priors
- parameter settings
- prior knowledge
- noise reduction
- lower bound
- metadata
- probabilistic model
- edge detection