Composite Embedding Systems Based On Dnn-Hmm And Attention End-To-End For Zerospeech2017 Track1 (2)

Shibata Hayato, Taku Kato, Shinozaki Takahiro, Shinji WatanabeZenodo 2017 · DOI · OpenAlex

Abstract

Deep neural networks (DNNs) were trained for posterior and bottleneck features using Japanese and other language speech data. We explore various DNN types, their combinations, and dimension reduction by principal component analysis (PCA). This version (version 2 ) concatenates CSJ feature vector and PCA compressed feature vector made from attention end-to-end feature. X:CSJ feature (60 dim bottleneck, (version 1 feature)) S:Attention end-to-end feature (320 dim) T:PCA(S) (60 dim) Z=concat(X,T)
BibTeX
@inproceedings{hayato2017composite,
  author = {Shibata Hayato and Taku Kato and Shinozaki Takahiro and Shinji Watanabe},
  title = {{Composite Embedding Systems Based On Dnn-Hmm And Attention End-To-End For Zerospeech2017 Track1 (2)}},
  booktitle = {Zenodo},
  year = {2017},
  doi = {10.5281/zenodo.823694},
}

Loading related content...

Figure