Exchangeability and Kernel Invariance in Trained MLPs

Open in new window