A Meta-Level Learning Algorithm for Sequential Hyper-Parameter Space Reduction in AutoML.
Giorgos BorboudakisPaulos CharonyktakisKonstantinos ParaschakisIoannis TsamardinosPublished in: CoRR (2023)
Keyphrases
- meta level
- space reduction
- learning algorithm
- hyperparameters
- object level
- random sampling
- cross validation
- model selection
- closed form
- prior information
- bayesian inference
- domain knowledge
- parameter settings
- active learning
- feature space
- principal component analysis
- em algorithm
- support vector
- supervised learning
- incremental learning
- training examples
- noise level
- reinforcement learning
- machine learning
- sample size
- incomplete data
- maximum a posteriori
- maximum likelihood
- training data
- face recognition
- low level
- bayesian framework
- high level
- prior knowledge
- object recognition
- pixel level
- computer vision