Header menu link for other important links
X
Generalized distillation framework for speaker normalization
Published in International Speech Communication Association
2017
Volume: 2017-August
   
Pages: 739 - 743
Abstract
Generalized distillation framework has been shown to be effective in speech enhancement in the past. We extend this idea to speaker normalization without any explicit adaptation data in this paper. In the generalized distillation framework, we assume the presence of some "privileged" information to guide the training process in addition to the training data. In the proposed approach, the privileged information is obtained from a "teacher" model, trained on speaker-normalized FMLLR features. The "student" model is trained on un-normalized filterbank features and uses teacher's supervision for cross-entropy training. The proposed distillation method does not need first pass decode information during testing and imposes no constraints on the duration of the test data for computing speakerspecific transforms unlike in FMLLR or i-vector. Experiments done on Switchboard and AMI corpus show that the generalized distillation framework shows improvement over un-normalized features with or without i-vectors. Copyright © 2017 ISCA.
About the journal
JournalProceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH
PublisherInternational Speech Communication Association
ISSN2308457X
Open AccessYes
Concepts (14)
  •  related image
    ELECTRIC SWITCHBOARDS
  •  related image
    Personnel training
  •  related image
    Speech communication
  •  related image
    Speech enhancement
  •  related image
    Teaching
  •  related image
    Cross entropy
  •  related image
    DISTILLATION METHOD
  •  related image
    FMLLR
  •  related image
    I VECTORS
  •  related image
    SPEAKER NORMALIZATION
  •  related image
    SWITCHBOARD
  •  related image
    Training data
  •  related image
    TRAINING PROCESS
  •  related image
    Distillation