A Gradient-based Bilevel Optimization Approach for Tuning Hyperparameters in Machine Learning.
Ankur SinhaTanmay KhandaitRaja MohantyPublished in: CoRR (2020)
Keyphrases
- hyperparameters
- machine learning
- model selection
- grid search
- parameter optimization
- cross validation
- bayesian inference
- bilevel programming
- support vector
- gaussian process
- bayesian framework
- parameter settings
- prior information
- support vector machine
- random sampling
- closed form
- maximum likelihood
- maximum a posteriori
- gaussian processes
- incremental learning
- incomplete data
- generalization ability
- data mining
- noise level
- decision trees
- computer vision
- random forest
- sample size
- em algorithm
- graph cuts
- learning algorithm
- probabilistic model
- feature space
- data analysis
- convex programming
- feature selection