Data-Dependent Stability of Stochastic Gradient Descent.
Ilja KuzborskijChristoph H. LampertPublished in: ICML (2018)
Keyphrases
- data dependent
- stochastic gradient descent
- least squares
- matrix factorization
- loss function
- step size
- online algorithms
- random forests
- support vector machine
- weight vector
- energy functional
- multiresolution
- importance sampling
- regularization parameter
- computer vision
- lower bound
- multiple kernel learning
- missing data
- active contours
- collaborative filtering
- training data