Anchor & Transform: Learning Sparse Embeddings for Large Vocabularies.

Paul Pu Liang, Manzil Zaheer, Yuan Wang, Amr Ahmed

Browse the full ICLR paper archive.