Joint Training of Expanded End-to-End DNN for Text-Dependent Speaker Verification

Heo, Hee-Soo; Jung, Jee-weon; Yang, Il-Hyung; Yoon, Sung-Hyun; Yu, Ha-Jin

doi:10.21437/interspeech.2017-1050

Cited by 24 publications

(17 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The batch size for training the two models was 40. For efficient training of the spectrogram-based CNN-GRU model, we trained the CNN part except the GRU layer in the whole model and then re-trained after attaching the GRU layer on the CNN, following the multi-step training scheme reported in [9,21]. Table 1 demonstrates the baseline of this study, which is an improved version of the authors' submission to DCASE 2018 competition.…”

Section: Experimental Settingsmentioning

confidence: 99%

Acoustic Scene Classification Using Teacher-Student Learning with Soft-Labels

et al. 2019

Self Cite

View full text Add to dashboard Cite

Acoustic scene classification identifies an input segment into one of the pre-defined classes using spectral information. The spectral information of acoustic scenes may not be mutually exclusive due to common acoustic properties across different classes, such as babble noises included in both airports and shopping malls. However, conventional training procedure based on one-hot labels does not consider the similarities between different acoustic scenes. We exploit teacher-student learning with the purpose to derive soft-labels that consider common acoustic properties among different acoustic scenes. In teacher-student learning, the teacher network produces softlabels, based on which the student network is trained. We investigate various methods to extract soft-labels that better represent similarities across different scenes. Such attempts include extracting soft-labels from multiple audio segments that are defined as an identical acoustic scene. Experimental results demonstrate the potential of our approach, showing a classification accuracy of 77.36 % on the DCASE 2018 task 1 validation set.

show abstract

Section: Experimental Settingsmentioning

confidence: 99%

Acoustic Scene Classification Using Teacher-Student Learning with Soft-Labels

et al. 2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…Deep networks often exploit pre-training schemes to show improved generalization performance. One such scheme was introduced by Heo et al [10]. This scheme trains the DNN over several stages, each stage using the parameters of preceding DNN as initialization.…”

Section: Multi-step Trainingmentioning

confidence: 99%

“…Studies on raw waveform processing and back-end classification have also occurred [3,6,9]. Individual DNNs are integrated to comprise end-to-end DNNs [1,3,10,11,12,13].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Avoiding Speaker Overfitting in End-to-End DNNs Using Raw Waveform for Text-Independent Speaker Verification

et al. 2018

Self Cite

View full text Add to dashboard Cite

In this research, we propose a novel raw waveform endto-end DNNs for text-independent speaker verification. For speaker verification, many studies utilize the speaker embedding scheme, which trains deep neural networks as speaker identifiers to extract speaker features. However, this scheme has an intrinsic limitation in which the speaker feature, trained to classify only known speakers, is required to represent the identity of unknown speakers. Owing to this mismatch, speaker embedding systems tend to well generalize towards unseen utterances from known speakers, but are overfitted to known speakers. This phenomenon is referred to as speaker overfitting. In this paper, we investigated regularization techniques, a multistep training scheme, and a residual connection with pooling layers in the perspective of mitigating speaker overfitting which lead to considerable performance improvements. Technique effectiveness is evaluated using the VoxCeleb dataset, which comprises over 1,200 speakers from various uncontrolled environments. To the best of our knowledge, we are the first to verify the success of end-to-end DNNs directly using raw waveforms in text-independent scenario. It shows an equal error rate of 7.4%, which is lower than i-vector/probabilistic linear discriminant analysis and end-to-end DNNs that use spectrograms.

show abstract

“…Recently, the speaker representation models have moved from the commonly used i-vector model [1,2,3], with a probabilistic linear discriminant (PLDA) back-end [4,5] to a new paradigm: speaker embedding trained from deep neural networks. Various speaker embeddings based on different network architectures [6,7] , attention mechanism [8,9], loss functions [10,11], noise robustness [12,13], and training paradigms [14,15] have been proposed and greatly improve the performance of speaker verification systems. Snyder et al [6] recently proposed the x-vector model, which is based on a Time-Delay Deep Neural Network (TDNN) architecture that computes speaker embeddings from variable-length acoustic segments.…”

Section: Introductionmentioning

confidence: 99%

Cross-lingual Text-independent Speaker Verification Using Unsupervised Adversarial Discriminative Domain Adaptation

Xia

Huang²,

Hansen

2019

ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Speaker verification systems often degrade significantly when there is a language mismatch between training and testing data. Being able to improve cross-lingual speaker verification system using unlabeled data can greatly increase the robustness of the system and reduce human labeling costs. In this study, we introduce an unsupervised Adversarial Discriminative Domain Adaptation (ADDA) method to effectively learn an asymmetric mapping that adapts the target domain encoder to the source domain, where the target domain and source domain are speech data from different languages. ADDA, together with a popular Domain Adversarial Training (DAT) approach, are evaluated on a cross-lingual speaker verification task: the training data is in English from NIST SRE04-08, Mixer 6 and Switchboard, and the test data is in Chinese from AISHELL-I. We show that with the ADDA adaptation, Equal Error Rate (EER) of the x-vector system decreases from 9.331% to 7.645%, relatively 18.07% reduction of EER, and 6.32% reduction from DAT as well. Further data analysis of ADDA adapted speaker embedding shows that the learned speaker embeddings can perform well on speaker classification for the target domain data, and are less dependent with respect to the shift in language.

show abstract

Joint Training of Expanded End-to-End DNN for Text-Dependent Speaker Verification

Cited by 24 publications

References 14 publications

Acoustic Scene Classification Using Teacher-Student Learning with Soft-Labels

Acoustic Scene Classification Using Teacher-Student Learning with Soft-Labels

Avoiding Speaker Overfitting in End-to-End DNNs Using Raw Waveform for Text-Independent Speaker Verification

Cross-lingual Text-independent Speaker Verification Using Unsupervised Adversarial Discriminative Domain Adaptation

Contact Info

Product

Resources

About