CONF
valente:Icslp:2006/IDIAP
Infinite Models for Speaker Clustering
Valente, Fabio
EXTERNAL
https://publications.idiap.ch/attachments/papers/2006/valente-Icslp-2006.pdf
PUBLIC
https://publications.idiap.ch/index.php/publications/showcite/valente:rr06-19
Related documents
International Conference on Spoken Language Processing
2006
IDIAP-RR 06-19
In this paper we propose the use of infinite models for the clustering of speakers. Speaker segmentation is obtained trough a Dirichlet Process Mixture (DPM) model which can be interpreted as a flexible model with an infinite a priori number of components. Learning is based on a Variational Bayesian approximation of the infinite sequence. DPM model is compared with fixed prior systems learned by ML/BIC, MAP/BIC and a Variational Bayesian method. Experiments are run on a speaker clustering task on the NIST-96 Broadcast News database.
REPORT
valente:rr06-19/IDIAP
Infinite Models for Speaker Clustering
Valente, Fabio
EXTERNAL
https://publications.idiap.ch/attachments/reports/2006/valente-idiap-rr-06-19.pdf
PUBLIC
Idiap-RR-19-2006
2006
IDIAP
Published in ICLSP 2006
In this paper we propose the use of infinite models for the clustering of speakers. Speaker segmentation is obtained trough a Dirichlet Process Mixture (DPM) model which can be interpreted as a flexible model with an infinite a priori number of components. Learning is based on a Variational Bayesian approximation of the infinite sequence. DPM model is compared with fixed prior systems learned by ML/BIC, MAP/BIC and a Variational Bayesian method. Experiments are run on a speaker clustering task on the NIST-96 Broadcast News database.