Interpretable global minima of deep ReLU neural networks on sequentially separable data

Open in new window