Detail publikace

Multilingually Trained Bottleneck Features in Spoken Language Recognition

FÉR, R. MATĚJKA, P. GRÉZL, F. PLCHOT, O. VESELÝ, K. ČERNOCKÝ, J.

Originální název

Multilingually Trained Bottleneck Features in Spoken Language Recognition

Anglický název

Multilingually Trained Bottleneck Features in Spoken Language Recognition

Jazyk

en

Originální abstrakt

Multilingual training of neural networks has proven to be simple yet effective way to deal with multilingual training corpora. It allows to use several resources to jointly train a language independent representation of features, which can be encoded into low-dimensional feature set by embedding narrow bottleneck layer to the network. In this paper, we analyze such features on the task of spoken language recognition (SLR), focusing on practical aspects of training bottleneck networks and analyzing their integration in SLR. By comparing properties of mono and multilingual features we show the suitability of multilingual training for SLR. The state-of-the-art performance of these features is demonstrated on the NIST LRE09 database.

Anglický abstrakt

Multilingual training of neural networks has proven to be simple yet effective way to deal with multilingual training corpora. It allows to use several resources to jointly train a language independent representation of features, which can be encoded into low-dimensional feature set by embedding narrow bottleneck layer to the network. In this paper, we analyze such features on the task of spoken language recognition (SLR), focusing on practical aspects of training bottleneck networks and analyzing their integration in SLR. By comparing properties of mono and multilingual features we show the suitability of multilingual training for SLR. The state-of-the-art performance of these features is demonstrated on the NIST LRE09 database.

Dokumenty

BibTex


@article{BUT144471,
  author="Radek {Fér} and Pavel {Matějka} and František {Grézl} and Oldřich {Plchot} and Karel {Veselý} and Jan {Černocký}",
  title="Multilingually Trained Bottleneck Features in Spoken Language Recognition",
  annote="Multilingual training of neural networks has proven to be simple yet effective
way to deal with multilingual training corpora. It allows to use several
resources to jointly train a language independent representation of features,
which can be encoded into low-dimensional feature set by embedding narrow
bottleneck layer to the network. In this paper, we analyze such features on the
task of spoken language recognition (SLR), focusing on practical aspects of
training bottleneck networks and analyzing their integration in SLR. By comparing
properties of mono and multilingual features we show the suitability of
multilingual training for SLR. The state-of-the-art performance of these features
is demonstrated on the NIST LRE09 database.",
  address="NEUVEDEN",
  chapter="144471",
  doi="10.1016/j.csl.2017.06.008",
  edition="NEUVEDEN",
  howpublished="online",
  institution="NEUVEDEN",
  number="46",
  volume="2017",
  year="2017",
  month="july",
  pages="252--267",
  publisher="NEUVEDEN",
  type="journal article in Web of Science"
}