First-order Stochastic Algorithms for Escaping From Saddle Points in Almost Linear Time

Open in new window