Dataset Open Access

Multilingual bottle-neck feature learning from untranscribed speech for track 1 in zerospeech2017 (system 2 -- with VTLN)

Hongjie Chen Chen; Cheung-Chi Leung; Lei Xie; Bin Ma; Haizhou Li


JSON-LD (schema.org) Export

{
  "description": "<p>We investigate the extraction of bottle-neck features (BNFs) for multiple languages without access to manual transcription.\u00a0Multilingual BNFs are derived from a multi-task learning deep neural network which is trained with unsupervised phoneme-like labels. The unsupervised phoneme-like labels are obtained from language-dependent Dirichlet process Gaussian mixture models separately trained on untranscribed speech of multiple languages.</p>\n\n<blockquote>\n<p>In this version, the input MFCC for DPGMM is processed with VTLN.</p>\n</blockquote>\n\n<p>\u00a0</p>", 
  "license": "http://creativecommons.org/licenses/by/4.0/legalcode", 
  "creator": [
    {
      "affiliation": "Northwestern Polytechnical University", 
      "@type": "Person", 
      "name": "Hongjie Chen Chen"
    }, 
    {
      "affiliation": "Institute for Infocomm Research, A*STAR", 
      "@type": "Person", 
      "name": "Cheung-Chi Leung"
    }, 
    {
      "affiliation": "Northwestern Polytechnical University", 
      "@type": "Person", 
      "name": "Lei Xie"
    }, 
    {
      "affiliation": "Institute for Infocomm Research, A*STAR", 
      "@type": "Person", 
      "name": "Bin Ma"
    }, 
    {
      "affiliation": "National University of Singapore", 
      "@type": "Person", 
      "name": "Haizhou Li"
    }
  ], 
  "url": "https://zenodo.org/record/822737", 
  "datePublished": "2017-07-04", 
  "@context": "https://schema.org/", 
  "distribution": [
    {
      "contentUrl": "https://zenodo.org/api/files/44b6cd5a-ff2b-4e99-954f-eee8e2d57d95/10_5281_zenodo_822737.tar.gz", 
      "@type": "DataDownload", 
      "fileFormat": "gz"
    }
  ], 
  "identifier": "https://doi.org/10.5281/zenodo.822737", 
  "@id": "https://doi.org/10.5281/zenodo.822737", 
  "@type": "Dataset", 
  "name": "Multilingual bottle-neck feature learning from untranscribed speech for track 1 in zerospeech2017 (system 2 -- with VTLN)"
}
86
29
views
downloads
All versions This version
Views 8686
Downloads 2929
Data volume 233.1 GB233.1 GB
Unique views 8383
Unique downloads 2424

Share

Cite as