Embeddings in Natural Language Processing