The Directional Bias Helps Stochastic Gradient Descent to Generalize in Kernel Regression Models.
Yiling LuoXiaoming HuoYajun MeiPublished in: CoRR (2022)
Keyphrases
- regression model
- stochastic gradient descent
- multiple kernel learning
- least squares
- loss function
- matrix factorization
- step size
- model selection
- kernel function
- kernel methods
- support vector machine
- random forests
- support vector
- regularization parameter
- linear model
- weight vector
- linear svm
- gaussian process
- feature space
- hyperparameters
- importance sampling
- cross validation
- online algorithms
- linear combination
- multi task
- random forest
- feature selection
- feature vectors
- data points
- feature set