Detail publikace

Language Recognition in iVectors Space

Originální název

Language Recognition in iVectors Space

Anglický název

Language Recognition in iVectors Space

Jazyk

en

Originální abstrakt

We have introduced a novel approach for language recognition. Three classifiers (linear generative model, SVM and logistic regression) have been tested in the iVector space, and all outperform the state-of-the-art JFA system. Very simple and fast classifier based on linear generative model provides excellent performance over all conditions. The advantage of this classifier is also its scalability: addition of a new language only requires estimating the mean over the corresponding iVectors. Most of the computational load is in the iVector generation. Hence, as a next step, we will try to obtain iVectors from the utterances and the corresponding sufficient statistics in a more direct way.

Anglický abstrakt

We have introduced a novel approach for language recognition. Three classifiers (linear generative model, SVM and logistic regression) have been tested in the iVector space, and all outperform the state-of-the-art JFA system. Very simple and fast classifier based on linear generative model provides excellent performance over all conditions. The advantage of this classifier is also its scalability: addition of a new language only requires estimating the mean over the corresponding iVectors. Most of the computational load is in the iVector generation. Hence, as a next step, we will try to obtain iVectors from the utterances and the corresponding sufficient statistics in a more direct way.

BibTex


@inproceedings{BUT76437,
  author="David {Martínez González} and Oldřich {Plchot} and Lukáš {Burget} and Ondřej {Glembek} and Pavel {Matějka}",
  title="Language Recognition in iVectors Space",
  annote="We have introduced a novel approach for language recognition. Three classifiers
(linear generative model, SVM and logistic regression) have been tested in the
iVector space, and all outperform the state-of-the-art JFA system. Very simple
and fast classifier based on linear generative model provides excellent
performance over all conditions. The advantage of this classifier is also its
scalability: addition of a new language only requires estimating the mean over
the corresponding iVectors. Most of the computational load is in the iVector
generation. Hence, as a next step, we will try to obtain iVectors from the
utterances and the corresponding sufficient statistics in a more direct way.",
  address="International Speech Communication Association",
  booktitle="Proceedings of Interspeech 2011",
  chapter="76437",
  edition="NEUVEDEN",
  howpublished="print",
  institution="International Speech Communication Association",
  number="8",
  year="2011",
  month="august",
  pages="861--864",
  publisher="International Speech Communication Association",
  type="conference paper"
}