Train faster, generalize better: Stability of stochastic gradient descent.
Moritz HardtBen RechtYoram SingerPublished in: ICML (2016)
Keyphrases
- stochastic gradient descent
- loss function
- least squares
- matrix factorization
- step size
- random forests
- online algorithms
- multiple kernel learning
- support vector machine
- regularization parameter
- importance sampling
- weight vector
- convergence rate
- linear svm
- feature extraction
- evolutionary algorithm
- pairwise
- image restoration
- particle swarm optimization
- data points
- support vector