High-risk learning: acquiring new word vectors from tiny data.

Aurlie Herbelot, Marco Baroni

Browse the full EMNLP paper archive.