Publication Details
Pairwise Discriminative Speaker Verification in the I -Vector Space
Brummer Johan Nikolaas Langenhoven (Agnitio)
Burget Lukáš, doc. Ing., Ph.D. (DCGM FIT BUT)
Laface Pietro, prof. (POLITO)
Plchot Oldřich, Ing., Ph.D. (DCGM FIT BUT)
Vasilakakis Vasileios, M.Sc. (POLITO)
Discriminative training, I-vector, large-scale training, probabilistic linear discriminant analysis, speaker recog- nition, support vector machines
In this work we present a novel framework for discriminative training of speaker verification systems, where a trial is represented, as in the PLDA approach, by an i-vector pair, and the task is discrimination between same-speaker and dif- ferent-speaker classes. This pairwise SVM approach provides a more natural paradigm to speaker verification compared to the classical one-vs-all discriminative training.
This work presents a new and efficient approach to discriminative speaker verification in the i-vector space. We illustrate the development of a linear discriminative classifier that is trained to discriminate between the hypothesis that a pair of feature vectors in a trial belong to the same speaker or to different speakers. This approach is alternative to the usual discriminative setup that discriminates between a speaker and all the other speakers. We use a discriminative classifier based on a Support Vector Machine (SVM) that is trained to estimate the parameters of a symmetric quadratic function approximating a log-likelihood ratio score without explicit modeling of the -vector distributions as in the generative Probabilistic Linear Discriminant Analysis (PLDA) models. Training these models is feasible because it is not necessary to expand the -vector pairs, which would be expensive or even impossible even for medium sized training sets. The results of experiments performed on the tel-tel extended core condition of the NIST 2010 Speaker Recognition Evaluation are competitive with the ones obtained by generative models, in terms of normalized Detection Cost Function and Equal Error Rate.Moreover, we show that it is possible to train a gender-independent discriminative model that achieves state-of-the-art accuracy, comparable to the one of a gender-dependent system, saving memory and execution time both in training and in testing.
@ARTICLE{FITPUB10450,
author = "Sandro Cumani and Langenhoven Nikolaas Johan Brummer and Luk\'{a}\v{s} Burget and Pietro Laface and Old\v{r}ich Plchot and Vasileios Vasilakakis",
title = "Pairwise Discriminative Speaker Verification in the I -Vector Space",
pages = "1217--1227",
journal = "IEEE Transactions on Audio, Speech, and Language Processing",
volume = 2013,
number = 6,
year = 2013,
ISSN = "1558-7916",
doi = "10.1109/TASL.2013.2245655",
language = "english",
url = "https://www.fit.vut.cz/research/publication/10450"
}