Implicit Hypersurface Approximation Capacity in Deep ReLU Networks

Open in new window