Abstract | ||
---|---|---|
We develop a class of algorithms, as variants of the stochastically controlled stochastic gradient (SCSG) methods [21], for the smooth non-convex finite-sum optimization problem. Assuming the smoothness of each component, the complexity of SCSG to reach a stationary point with E parallel to del f(x)parallel to(2) <= epsilon is O (min{epsilon(-5/3), epsilon(-1)n(2/3)}), which strictly outperforms the stochastic gradient descent. Moreover, SCSG is never worse than the state-of-the-art methods based on variance reduction and it significantly outperforms them when the target accuracy is low. A similar acceleration is also achieved when the functions satisfy the Polyak-Lojasiewicz condition. Empirical experiments demonstrate that SCSG outperforms stochastic gradient methods on training multi-layers neural networks in terms of both training and validation loss. |
Year | Venue | DocType |
---|---|---|
2017 | ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 30 (NIPS 2017) | Conference |
Volume | ISSN | Citations |
30 | 1049-5258 | 14 |
PageRank | References | Authors |
0.53 | 15 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Lihua Lei | 1 | 24 | 5.52 |
Ju, Cheng | 2 | 15 | 1.22 |
Jianbo Chen | 3 | 82 | 4.35 |
Michael I. Jordan | 4 | 31220 | 3640.80 |