Stagewise Training Accelerates Convergence of Testing Error Over SGD.
Zhuoning YuanYan YanRong JinTianbao YangPublished in: NeurIPS (2019)
Keyphrases
- stochastic gradient descent
- test set
- testing phase
- training set
- training stage
- training process
- training algorithm
- supervised learning
- loss function
- convergence speed
- online learning
- error rate
- error bounds
- empirical risk
- risk minimization
- software testing
- generalization error
- matrix factorization
- data sets
- test data
- least squares
- pairwise
- lower bound
- neural network