CP-SGD: Distributed stochastic gradient descent with compression and periodic compensation.
Enda YuDezun DongYemao XuShuo OuyangXiangke LiaoPublished in: J. Parallel Distributed Comput. (2022)
Keyphrases
- stochastic gradient descent
- least squares
- loss function
- step size
- matrix factorization
- stochastic gradient
- random forests
- weight vector
- online algorithms
- multiple kernel learning
- alternating least squares
- regularization parameter
- image compression
- support vector machine
- importance sampling
- convergence speed
- cost function
- negative matrix factorization
- logistic regression
- worst case