On the Impact of Overparameterization on the Training of a Shallow Neural Network in High Dimensions

Open in new window