Passer à la navigation principale Passer à la recherche Passer au contenu principal

Maximum likelihood estimation of eigenvoices and residual variances for large vocabulary speech recognition tasks

  • Computer Research Institute of Montreal

Résultats de recherche: Contribution à une conférence › Compte-rendu de conférence › Revue par des pairs

2 Citations (Scopus)

Résumé

We describe a new algorithm for estimating eigenvoices (or, equivalently, EMAP correlation matrices) for large vocabulary speech recognition tasks. The algorithm is an EM procedure based on a novel maximum likelihood formulation of the estimation problem which is similar to the mathematical model underlying probabilistic principal components analysis. It enables us to extend eigenvoice/EMAP adaptation in a natural way to adapt variances as well as mean vectors. It differs from other approaches in that it does not require that speaker dependent or speaker adapted models for the training speakers be given in advance (these are derived as a byproduct of the estimation procedure). Accordingly our algorithm can be applied directly to large vocabulary tasks even if the training data is sparse in the sense that only a small fraction of the total number of Gaussians is observed for each training speaker.

Langue originaleAnglais
Pages57-60
Nombre de pages4
ÉtatPublié - 2002
Modification externeOui
Evénement7th International Conference on Spoken Language Processing, ICSLP 2002 - Denver, Etats-Unis
Durée: 16 sept. 2002 → 20 sept. 2002

Conférence

Conférence7th International Conference on Spoken Language Processing, ICSLP 2002
Pays/TerritoireEtats-Unis
La villeDenver
Période16/09/02 → 20/09/02

Classification disciplinaire ASJC Scopus

  • Langues et linguistique
  • Linguistique et langues

Empreinte digitale

Voici les principaux termes ou expressions associés à « Maximum likelihood estimation of eigenvoices and residual variances for large vocabulary speech recognition tasks ». Ces libellés thématiques sont générés à partir du titre et du résumé de la publication. Ensemble, ils forment une empreinte digitale unique.

Citer cette ressource