Symmetric Rank-One Quasi-Newton Methods for Deep Learning Using Cubic Regularization

Open in new window