Conference paper

LOZANO-DIEZ Alicia, PLCHOT Oldřich, MATĚJKA Pavel, NOVOTNÝ Ondřej and GONZALEZ-RODRIGUEZ Joaquin. Analysis of DNN-based Embeddings for Language Recognition on the NIST LRE 2017. In: Proceedings of Odyssey 2018 The Speaker and Language Recognition Workshop. Les Sables d'Olonne: International Speech Communication Association, 2018, pp. 39-46. ISSN 2312-2846.
Publication language:english
Original title:Analysis of DNN-based Embeddings for Language Recognition on the NIST LRE 2017
Title (cs):Analýza DNN Embeddings pro rozpoznávání jazyka v NIST LRE 2017
Pages:39-46
Proceedings:Proceedings of Odyssey 2018 The Speaker and Language Recognition Workshop
Conference:Odyssey 2018
Place:Les Sables d'Olonne, FR
Year:2018
Journal:Proceedings of Odyssey: The Speaker and Language Recognition Workshop, Vol. 2018, No. 6, 4 Rue des Fauvettes - Lous Tourils, F-66390 BAIXAS, FR
ISSN:2312-2846
DOI:10.21437/Odyssey.2018
Publisher:International Speech Communication Association
URL:http://www.fit.vutbr.cz/research/groups/speech/publi/2018/lozano_odyssey2018_42.pdf [PDF]
Keywords
language recognition
Annotation
In this work, we analyze different designs of a language identification (LID) system based on embeddings. In our case, an embedding represents a whole utterance (or a speech segment of variable duration) as a fixed-length vector (similar to the ivector). Moreover, this embedding aims to capture information relevant to the target task (LID), and it is obtained by training a deep neural network (DNN) to classify languages. In particular, we trained a DNN based on bidirectional long short-term memory (BLSTM) recurrent neural network (RNN) layers, whose frame-by-frame outputs are summarized into mean and standard deviation statistics for each utterance. After this pooling layer, we add two fully connected layers whose outputs are used as embeddings, which are afterwards modeled by a Gaussian linear classifier (GLC). For training, we add a softmax output layer and train the whole network with multi-class cross-entropy objective to discriminate between languages. We analyze the effect of using data augmentation in the DNN training, as well as different input features and architecture hyper-parameters, obtaining configurations that gradually improved the performance of the embedding system. We report our results on the NIST LRE 2017 evaluation dataset and compare the performance of embeddings with a reference i-vector system. We show that the best configuration of our embedding system outperforms the strong reference i-vector system by 3% relative, and this is further pushed up to 10% relative improvement via a simple score level fusion.
BibTeX:
@INPROCEEDINGS{
   author = {Alicia Lozano-Diez and Old{\v{r}}ich Plchot and Pavel
	Mat{\v{e}}jka and Ond{\v{r}}ej Novotn{\'{y}} and Joaquin
	Gonzalez-Rodriguez},
   title = {Analysis of DNN-based Embeddings for Language Recognition on
	the NIST LRE 2017},
   pages = {39--46},
   booktitle = {Proceedings of Odyssey 2018 The Speaker and Language
	Recognition Workshop},
   journal = {Proceedings of Odyssey: The Speaker and Language Recognition
	Workshop},
   volume = {2018},
   number = {6},
   year = {2018},
   location = {Les Sables d'Olonne, FR},
   publisher = {International Speech Communication Association},
   ISSN = {2312-2846},
   doi = {10.21437/Odyssey.2018},
   language = {english},
   url = {http://www.fit.vutbr.cz/research/view_pub.php?id=11761}
}

Your IPv4 address: 54.225.38.2
Switch to IPv6 connection

DNSSEC [dnssec]