Towards a Mathematical Understanding of the Difficulty in Learning with Feedforward Neural Networks

Open in new window