OL\'E: Orthogonal Low-rank Embedding, A Plug and Play Geometric Loss for Deep Learning

Open in new window