Scaling Limit of Neural Networks with the Xavier Initialization and Convergence to a Global Minimum

Open in new window