Using Data-driven and Phonetic Units for Speaker Verification | Publicación