Analysis of BUT-PT Submission for NIST LRE 2017

Plchot, Oldřich; Matějka, Pavel; Novotny, Ondrej; Cumani, Sandro; Lozano-Díez, Alicia; Slavíček, Josef; Díez, Mireia; Grézl, František; Glembek, Ondřej; Kamsali, Mounika; Silnova, Anna; Burget, Lukáš; Ondel, Lucas; Kesiraju, Santosh; Rohdin, Johan

doi:10.21437/odyssey.2018-7

Cited by 9 publications

(10 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In these cases, the performance is even slightly degraded. We have also performed a similar analysis for the i-vector system, where we did not see any benefits from data augmentation [24].…”

Section: Experiments On Data Augmentation For Embeddingsmentioning

confidence: 99%

“…In this work, we continue the research started in [22] while exploring and analyzing DNN embeddings for language recognition on the most recent NIST LRE 2017. In particular, we describe the embedding subsystem submitted to this last NIST LRE [24], and analyze further improvements made after the evaluation period. We study the influence of increasing the size of our original system architecture, as well as the performance when using different input features and a wide range of augmented training data.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Analysis of DNN-based Embeddings for Language Recognition on the NIST LRE 2017

Lozano-Díez¹,

Plchot²,

Matějka³

et al. 2018

The Speaker and Language Recognition Workshop (Odyssey 2018)

Self Cite

View full text Add to dashboard Cite

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.

show abstract

Section: Experiments On Data Augmentation For Embeddingsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Analysis of DNN-based Embeddings for Language Recognition on the NIST LRE 2017

Lozano-Díez¹,

Plchot²,

Matějka³

et al. 2018

The Speaker and Language Recognition Workshop (Odyssey 2018)

Self Cite

View full text Add to dashboard Cite

show abstract

“…VAD-NN The NN which produces per-frame posterior probabilities for speech and non-speech classes that are later post-processed to create continuous speech segments was trained on the 8kHz Fisher English [13].…”

Section: Voice Activity Detectionmentioning

confidence: 99%

Analysis of BUT Submission in Far-Field Scenarios of VOiCES 2019 Challenge

et al. 2019

Self Cite

View full text Add to dashboard Cite

This paper is a post-evaluation analysis of our efforts in VOiCES 2019 Speaker Recognition challenge. All systems in the fixed condition are based on x-vectors with different features and DNN topologies. The single best system reaches minDCF of 0.38 (5.25% EER) and a fusion of 3 systems yields minDCF of 0.34 (4.87% EER). We also analyze how speaker verification (SV) systems evolved in last few years and show results also on SITW 2016 Challenge. EER on the core-core condition of the SITW 2016 challenge dropped from 5.85% to 1.65% for system fusions submitted for SITW 2016 and VOiCES 2019, respectively. The less restrictive open condition allowed us to use external data for PLDA adaptation and achieve additional small performance improvement. In our submission to open condition, we used three x-vector systems and also one system based on i-vectors.

show abstract

“…A thorough summary of our recent work with BN features in LID has been presented in [9]. Finally, SBN features again contributed to a superior performance of BUT-UAM-Phonexia-PoliTo system in 2017 NIST Language recognition evaluation (LRE17) [10].…”

Section: Introductionmentioning

confidence: 98%

“…2, ii) details on the training of included NNs in Sec. 3, and iii) a summary of results achieved with BN features extracted with the released package on NIST LRE15 and LRE17 data [10,16]. Sec.…”

Section: Introductionmentioning

confidence: 99%

BUT/Phonexia Bottleneck Feature Extractor

Silnova¹,

Matějka²,

Glembek³

et al. 2018

The Speaker and Language Recognition Workshop (Odyssey 2018)

Self Cite

View full text Add to dashboard Cite

This paper complements the public release of the BUT/Phonexia bottleneck (BN) feature extractor. Starting with a brief history of Neural Network (NN)-based and BN-based approaches to speech feature extraction, it describes the structure of the released software. It follows by describing the three provided NNs: the first two trained on the US English Fisher corpus with monophone-state and tied-state targets, and the third network trained in a multilingual fashion on 17 Babel languages. The NNs were technically trained to classify acoustic units, however the networks were optimized with respect to the language recognition task, which is the main focus of this paper. Nevertheless, it is worth noting that apart from language recognition, the provided software can be used with any speech-related task. The paper concludes with a comprehensive summary of the results obtained on the NIST 2015 and 2017 Language Recognition Evaluations tasks.

show abstract

Analysis of BUT-PT Submission for NIST LRE 2017

Cited by 9 publications

References 0 publications

Analysis of DNN-based Embeddings for Language Recognition on the NIST LRE 2017

Analysis of DNN-based Embeddings for Language Recognition on the NIST LRE 2017

Analysis of BUT Submission in Far-Field Scenarios of VOiCES 2019 Challenge

BUT/Phonexia Bottleneck Feature Extractor

Contact Info

Product

Resources

About