S-SGD: Symmetrical Stochastic Gradient Descent with Weight Noise Injection for Reaching Flat Minima.
Wonyong SungIksoo ChoiJinhwan ParkSeokhyun ChoiSungho ShinPublished in: CoRR (2020)
Keyphrases
- stochastic gradient descent
- weight vector
- least squares
- loss function
- matrix factorization
- step size
- stochastic gradient
- random forests
- multiple kernel learning
- noise level
- missing data
- online algorithms
- alternating least squares
- support vector machine
- linear combination
- regularization parameter
- learning rate
- cost function
- feature extraction
- importance sampling
- hyperplane
- feature selection
- logistic regression
- online learning
- decision trees