Tied Mixture of Factor Analyzers Layer to Combine Frame Level Representations in Neural Speaker Embeddings.

Nanxin Chen, Jess Villalba, Najim Dehak

Browse the full Interspeech paper archive.