An Improved Analysis of Training Over-parameterized Deep Neural Networks

Open in new window