Глубокий канонический корреляционный анализ

Word embeddings have been found useful for many NLP tasks, including part-of-speech tagging, named entity recognition, and pars- ing. Adding multilingual context when learn- ing embeddings can improve their quality, …...