Login / Signup

Investigations on features for log-linear acoustic models in continuous speech recognition.

Simon WieslerMarkus Nußbaum-ThomGeorg HeigoldRalf SchlüterHermann Ney
Published in: ASRU (2009)
Keyphrases
  • continuous speech recognition
  • log linear
  • log linear models
  • low level
  • feature extraction
  • prior knowledge
  • co occurrence
  • generative model
  • unsupervised learning