Semi-supervised Nuisance-attribute Networks for Domain Adaptation

Lin, Weiwei; Mak, Man‐Wai; Tu, Youzhi; Chien, Jen‐Tzung

doi:10.1109/icassp.2019.8682752

Cited by 5 publications

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [264], Lin et al added a MMD based loss to the reconstruction loss of an autoencoder to train a domain-invariant encoder for multi-source adaptation of ivectors. In [265,266], they further proposed a nuisanceattribute autoencoder based on MMD. In [267], they proposed a multi-level deep neural network adaptation method using MMD and consistency regularization.…”

Section: Discrepancy-based Domain Adaptationmentioning

confidence: 99%

Speaker Recognition Based on Deep Learning: An Overview

Bai

Zhang

2020

Preprint

View full text Add to dashboard Cite

Speaker recognition is a task of identifying persons from their voices. Recently, deep learning has dramatically revolutionized speaker recognition. However, there is lack of comprehensive reviews on the exciting progress. In this paper, we review several major subtasks of speaker recognition, including speaker verification, identification, diarization, and robust speaker recognition, with a focus on deep-learning-based methods. Because the major advantage of deep learning over conventional methods is its representation ability, which is able to produce highly abstract embedding features from utterances, we first pay close attention to deep-learning-based speaker feature extraction, including the inputs, network structures, temporal pooling strategies, and objective functions respectively, which are the fundamental components of many speaker recognition subtasks. Then, we make an overview of speaker diarization, with an emphasis of recent supervised, end-to-end, and online diarization. Finally, we survey robust speaker recognition from the perspectives of domain adaptation and speech enhancement, which are two major approaches of dealing with domain mismatch and noise problems. Popular and recently released corpora are listed at the end of the paper.

show abstract

Section: Discrepancy-based Domain Adaptationmentioning

confidence: 99%

Speaker Recognition Based on Deep Learning: An Overview

Bai

Zhang

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…In the recent NIST SRE challenges, one of the main interests was a language mismatch. To alleviate the language mismatch problem, several domain adaptation techniques were recently proposed [6,7,8,9,10]. In [6], an adversarial method for unsupervised discriminative domain adaptation was proposed.…”

Section: Introductionmentioning

confidence: 99%

“…In [6], an adversarial method for unsupervised discriminative domain adaptation was proposed. For reducing the domain mismatch in i-vector and x-vector SV systems, semi-supervised nuisance attribute network (SNAN) was introduced in [7]. Instead of computing the domain variability from the dataset means, maximum mean discrepancy (MMD) was used as part of the loss function.…”

Section: Introductionmentioning

confidence: 99%

Supervised Domain Adaptation for Text-Independent Speaker Verification Using Limited Data

et al. 2020

View full text Add to dashboard Cite

To adapt the speaker verification (SV) system to a target domain with limited data, this paper investigates the transfer learning of the model pre-trained on the source domain data. To that end, layer-by-layer adaptation with transfer learning from the initial and final layers of the pre-trained model is investigated. We show that the model adapted from the initial layers outperforms the model adapted from the final layers. Based on this evidence, and inspired by the works in image recognition field, we hypothesize that low-level convolutional neural network (CNN) layers characterize domain-specific component while high-level CNN layers are domain-independent and have more discriminative power. For adapting these domain-specific components, angular margin softmax (AMSoftmax) applied on the CNN-based implementation of the x-vector architecture. In addition, to reduce the problem of over-fitting on the limited target data, transfer learning on the batch norm layers is investigated. Mean shift and covariance estimation of batch norm allows to map the represented components of the target domain to the source domain. Using TDNN and E-TDNN versions of the x-vectors as baseline models, the adapted models on the development set of NIST SRE 2018 outperformed the baselines with relative improvements of 11.0 and 13.8 %, respectively.

show abstract

“…In [11], autoencoderbased domain adaptation was proposed to transfer channel information from the source domain to the target domain. In [12,13], Lin et al applied maximum mean discrepancy to measure the degree of domain mismatch across multiple domains and incorporated the measure into the objective function for training autoencoders. The bottleneck features extracted from the autoencoders are shown to be less domain dependent, resulting in performance gain in SRE16 data.…”

Section: Introductionmentioning

confidence: 99%

Variational Domain Adversarial Learning for Speaker Verification

Mak

Chien³

2019

Interspeech 2019

Self Cite

View full text Add to dashboard Cite

Domain mismatch refers to the problem in which the distribution of training data differs from that of the test data. This paper proposes a variational domain adversarial neural network (VDANN), which consists of a variational autoencoder (VAE) and a domain adversarial neural network (DANN), to reduce domain mismatch. The DANN part aims to retain speaker identity information and learn a feature space that is robust against domain mismatch, while the VAE part is to impose variational regularization on the learned features so that they follow a Gaussian distribution. Thus, the representation produced by VDANN is not only speaker discriminative and domaininvariant but also Gaussian distributed, which is essential for the standard PLDA backend. Experiments on both SRE16 and SRE18-CMN2 show that VDANN outperforms the Kaldi baseline and the standard DANN. The results also suggest that VAE regularization is effective for domain adaptation.

show abstract

Semi-supervised Nuisance-attribute Networks for Domain Adaptation

Cited by 5 publications

References 20 publications

Speaker Recognition Based on Deep Learning: An Overview

Speaker Recognition Based on Deep Learning: An Overview

Supervised Domain Adaptation for Text-Independent Speaker Verification Using Limited Data

Variational Domain Adversarial Learning for Speaker Verification

Contact Info

Product

Resources

About