Differentiable Supervector Extraction for Encoding Speaker and Phrase Information in Text Dependent Speaker Verification | Publicación