Department of Computer Graphics and Multimedia

Report

NOVOTNÝ Ondřej, MATĚJKA Pavel, GLEMBEK Ondřej, PLCHOT Oldřich, GRÉZL František, BURGET Lukáš and ČERNOCKÝ Jan. DNN-based SRE Systems in Multi-Language Conditions. Brno: Faculty of Information Technology BUT, 2016. Available from: http://www.fit.vutbr.cz/research/groups/speech/publi/2016/dnn-based-sre_TECH_REP_v0.pdf
Publication language:english
Original title:DNN-based SRE Systems in Multi-Language Conditions
Title (cs):Systémy pro rozpoznávání mluvčího založené na DNN v multilingválních podmínkách
Pages:5
Place:Brno, CZ
Year:2016
URL:http://www.fit.vutbr.cz/research/groups/speech/publi/2016/dnn-based-sre_TECH_REP_v0.pdf
Publisher:Faculty of Information Technology BUT
URL:http://www.fit.vutbr.cz/research/groups/speech/publi/2016/dnn-based-sre_TECH_REP_v0.pdf [PDF]
Keywords
speaker recognition, multilinguality, DNN
Annotation
This work studies the usage of the (currently state-of-the-art) Deep Neural Networks (DNN) i-vector/PLDA-based speaker recognition systems in multi-language (especially non-English) conditions. On the ``Language Pack'' of the PRISM set, we evaluate the systems' performance using NIST's standard metrics. We study the use of multi-lingual DNN in place of the original English DNN on these multi-language conditions. We show that not only the gain from using DNNs vanishes, but also the DNN-based systems tend to produce de-calibrated scores under the studied conditions. This work gives suggestions for directions of future research rather than any particular solutions.
Abstract
This work studies the usage of the (currently state-of-the-art) Deep Neural Networks (DNN) i-vector/PLDA-based speaker recognition sys- tems in multi-language (especially non-English) conditions. On the "Lan- guage Pack" of the PRISM set, we evaluate the systems performance using NISTs standard metrics. We study the use of multi-lingual DNN in place of the original English DNN on these multi-language conditions. We show that not only the gain from using DNNs vanishes, but also the DNN-based systems tend to produce de-calibrated scores under the studied conditions. This work gives suggestions for directions of future research rather than any particular solutions.
BibTeX:
@TECHREPORT{
   author = {Ond{\v{r}}ej Novotn{\'{y}} and Pavel Mat{\v{e}}jka and
	Ond{\v{r}}ej Glembek and Old{\v{r}}ich Plchot and
	Franti{\v{s}}ek Gr{\'{e}}zl and Luk{\'{a}}{\v{s}} Burget and
	Jan {\v{C}}ernock{\'{y}}},
   title = {DNN-based SRE Systems in Multi-Language Conditions},
   pages = {5},
   year = {2016},
   location = {Brno, CZ},
   publisher = {Faculty of Information Technology BUT},
   language = {english},
   url = {http://www.fit.vutbr.cz/research/view_pub.php.en.iso-8859-2?id=11235}
}

Your IPv4 address: 54.92.163.188
Switch to IPv6 connection

DNSSEC [dnssec]