Login / Signup

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

Nanxin ChenJesús VillalbaNajim Dehak
Published in: INTERSPEECH (2019)
Keyphrases
  • higher level
  • distributed representations
  • lower level
  • neural network
  • feature selection
  • vector space
  • manifold learning
  • multi layer
  • speaker verification