Implicit Bias of Gradient Descent on Linear Convolutional Networks

Open in new window