Search

In this work, we tackle the problem of end-to-end language identification from speech. To this end, we propose the use of a residual convolutional neural network aiming at exploiting the ability of such architectures to take into account large contextual segments of input data. Moreover, in order for variable input lengths to be supported by the proposed setting, a self-attention mechanism is employed on top of the final convolutional layer. This results in a learnable temporal feature pooling scheme that allows for embedding varying duration utterances into a fixed dimension space. Evaluation is performed on data containing ten oriental languages under different test conditions, namely: short-duration recordings, confusing languages trials, as well as a set of trials in which non-target unseen languages are included. End-to-end evaluation of the proposed framework is thus shown to significantly outperform well-known benchmark methods under considered evaluation conditions.

In this article we propose a novel approach for adapting speaker embeddings to new domains based on adversarial training of neural networks. We apply our embeddings to the task of text-independent speaker verification, a challenging, real-world problem in biometric security. We further the development of end-to-end speaker embedding models by combing a novel 1-dimensional, self-attentive residual network, an angular margin loss function and adversarial training strategy. Our model is able to learn extremely compact, 64-dimensional speaker embeddings that deliver competitive performance on a number of popular datasets using simple cosine distance scoring. One the NIST-SRE 2016 task we are able to beat a strong i-vector baseline, while on the Speakers in the Wild task our model was able to outperform both i-vector and x-vector baselines, showing an absolute improvement of 2.19% over the latter. Additionally, we show that the integration of adversarial training consistently leads to a significant improvement over an unadapted model., Submitted to ICASSP 2019

This article presents a novel approach for learning domain-invariant speaker embeddings using Generative Adversarial Networks. The main idea is to confuse a domain discriminator so that is can't tell if embeddings are from the source or target domains. We train several GAN variants using our proposed framework and apply them to the speaker verification task. On the challenging NIST-SRE 2016 dataset, we are able to match the performance of a strong baseline x-vector system. In contrast to the the baseline systems which are dependent on dimensionality reduction (LDA) and an external classifier (PLDA), our proposed speaker embeddings can be scored using simple cosine distance. This is achieved by optimizing our models end-to-end, using an angular margin loss function. Furthermore, we are able to significantly boost verification performance by averaging our different GAN models at the score level, achieving a relative improvement of 7.2% over the baseline., Comment: Submitted to ICASSP 2019

Classification accuracy of the kNN algorithm is found to be adversely affected by the presence of outliers in the experimental datasets. An outlier score based on rank difference can be assigned to the points in these datasets by taking into consideration the distance and density of their local neighborhood points. In the present work, we introduce a generalized outlier informative distance measure where a factor based on the above score is used to modulate any potential distance function. Properties of the new outlier informative distance measure are presented. Experiments on several numeric datasets in the UCI machine learning repository clearly reveal the effectiveness of the proposed formulation.

We propose to improve the performance of i-vector based speaker verification by processing the i-vectors with a deep neural network before they are fed to a cosine distance or probabilistic linear discriminant analysis (PLDA) classifier. To this end we build on an existing model that we refer to as Non-linear Within Class Normalization (NWCN) and introduce a novel Speaker Classifier Network (SCN). Both models deliver impressive speaker verification performance, showing a 56% and 68% relative improvement over standard i-vectors when combined with a cosine distance backend. The NWCN model also reduces the equal error rate for PLDA from 1.78% to 1.63%. We also test these models under the constraints of domain mismatch, i.e. when no in-domain training data is available. Under these conditions, SCN features in combination with cosine distance performs better than the PLDA baseline, achieving an equal error rate of 2.92% as compared to 3.37%.

The automatic speaker verification spoofing and countermeasures challenge 2015 provides a common framework for the evaluation of spoofing countermeasures or anti-spoofing techniques in the presence of various seen and unseen spoofing attacks. This contribution proposes a system consisting of amplitude, phase, linear prediction residual, and combined amplitude - phase-based countermeasures for the detection of spoofing attacks. In this task we use following features: Mel-frequency cepstral coefficients (MFCC), product spectrum-based cepstral coefficients, modified group delay cepstral coefficients, weighted linear prediction group delay cepstral coefficients, linear prediction residual cepstral coefficients, cosine normalized phase-based cepstral features (CNPCC), and a combination of MFCC-CNPCC. The product spectrum-based features are influenced by both the amplitude and phase spectra. The Gaussian Mixture Model (GMM) classifier is used for the discrimination of the human and spoofed speech signals. Our primary submitted system is a linear fusion of the sub-systems based on the features mentioned above with fusion weights trained on the development dataset. Experimental results on the challenge evaluation data provided an average EER (equal error rate) of 0.041%, 5.347%, and 2.69% on the known, unknown and all (known + unknown) spoofing attacks, respectively. Among all the systems product spectrum-based cepstral coefficients- and conventional MFCC (without any feature normalization)based systems performed the best in terms of EER measure. On the known, unknown and all conditions the EER obtained by the MFCC and product spectrum-based features are 0.78% & 0.65%, 5.39% & 5.37% and 3.09% & 3.01%, respectively.