Leveraging Speaker Embeddings in End-to-End Neural Diarization for Two-Speaker Scenarios | Publicación