On the Benefits of Large Learning Rates for Kernel Methods.
Gaspard BeugnotJulien MairalAlessandro RudiPublished in: COLT (2022)
Keyphrases
- kernel methods
- learning rate
- gaussian kernels
- reproducing kernel hilbert space
- uniform convergence
- kernel function
- convergence rate
- gaussian kernel
- learning problems
- machine learning
- support vector
- learning algorithm
- feature space
- kernel matrix
- support vector machine
- convergence speed
- kernel pca
- high dimensional feature space
- graph kernels
- kernel principal component analysis
- learning tasks
- multiple kernel learning
- kernel machines
- data sets
- input space
- kernel learning
- real valued
- data mining