How to represent a word and predict it, too: improving tied architectures for language modelling.

Kristina Gulordava, Laura Aina, Gemma Boleda

Browse the full EMNLP paper archive.