66 citations · 100 across the 2 of their papers we have counts for
Showing cs.LGShow all
2 papers · 1 filter
cs.LG2016★ 66 cited
The Power of Normalization: Faster Evasion of Saddle Points
Kfir Y. Levy
A commonly used heuristic in non-convex optimization is Normalized Gradient Descent (NGD) - a variant of gradient descent in which only the direction of the gradient is taken into…
cs.LG2014★ 34 cited
Logistic Regression: Tight Bounds for Stochastic and Online Optimization
Elad Hazan, Tomer Koren, Kfir Y. Levy
The logistic loss function is often advocated in machine learning and statistics as a smooth and strictly convex surrogate for the 0-1 loss. In this paper we investigate the questi…