Learning Parallel Sentence Embeddings from Multilingual Data - 42Papers