eduzhai > Applied Sciences > Engineering >

Compact Speaker Embedding lrx-vector

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 5 pages

Abstract: Deep neural networks (DNN) have recently been widely used in speakerrecognition systems, achieving state-of-the-art performance on variousbenchmarks. The x-vector architecture is especially popular in this researchcommunity, due to its excellent performance and manageable computationalcomplexity. In this paper, we present the lrx-vector system, which is thelow-rank factorized version of the x-vector embedding network. The primaryobjective of this topology is to further reduce the memory requirement of thespeaker recognition system. We discuss the deployment of knowledge distillationfor training the lrx-vector system and compare against low-rank factorizationwith SVD. On the VOiCES 2019 far-field corpus we were able to reduce theweights by 28 compared to the full-rank x-vector system while keeping therecognition rate constant (1.83 EER).

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×