Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

Unsupervised Domain Adaptation via Domain Adversarial Training for Speaker Recognition

Abstract: 

The i-vector approach to speaker recognition has achieved good performance when the domain of the evaluation dataset is similar to that of the training dataset. However, in real-world applications, there is always a mismatch between the training and evaluation datasets, that leads to performance degradation. To address this problem, this paper proposes to learn the domain-invariant and speaker-discriminative speech representations via domain adversarial training. Specifically, with domain adversarial training method, we use a gradient reversal layer to remove the domain variation and project the different domain data into the same subspace. Moreover, we compare the proposed method with other state-of-the-art unsupervised domain adaptation techniques for the i-vector approach to speaker recognition (e.g. autoencoder based domain adaptation, inter dataset variability compensation, dataset-invariant covariance normalization, and so on).
Experiments on 2013 domain adaptation challenge (DAC) dataset demonstrate that the proposed method is not only effective in solving the dataset mismatch problem but also outperforms the compared unsupervised domain adaptation methods.

up
0 users have voted:

Paper Details

Authors:
Qing Wang, Wei Rao, Sining Sun, Lei Xie, Eng Siong Chng, Haizhou Li
Submitted On:
25 April 2018 - 2:23am
Short Link:
Type:
Presentation Slides
Event:
Presenter's Name:
Qing Wang
Paper Code:
SP-L5.5
Document Year:
2018
Cite

Document Files

icassp2018_slides_qingwang_Unsupervised Domain Adaptation via Domain Adversarial Training for Speaker Recognition.pdf

(65 downloads)

Subscribe

[1] Qing Wang, Wei Rao, Sining Sun, Lei Xie, Eng Siong Chng, Haizhou Li, "Unsupervised Domain Adaptation via Domain Adversarial Training for Speaker Recognition", IEEE SigPort, 2018. [Online]. Available: http://sigport.org/3173. Accessed: Jul. 20, 2018.
@article{3173-18,
url = {http://sigport.org/3173},
author = {Qing Wang; Wei Rao; Sining Sun; Lei Xie; Eng Siong Chng; Haizhou Li },
publisher = {IEEE SigPort},
title = {Unsupervised Domain Adaptation via Domain Adversarial Training for Speaker Recognition},
year = {2018} }
TY - EJOUR
T1 - Unsupervised Domain Adaptation via Domain Adversarial Training for Speaker Recognition
AU - Qing Wang; Wei Rao; Sining Sun; Lei Xie; Eng Siong Chng; Haizhou Li
PY - 2018
PB - IEEE SigPort
UR - http://sigport.org/3173
ER -
Qing Wang, Wei Rao, Sining Sun, Lei Xie, Eng Siong Chng, Haizhou Li. (2018). Unsupervised Domain Adaptation via Domain Adversarial Training for Speaker Recognition. IEEE SigPort. http://sigport.org/3173
Qing Wang, Wei Rao, Sining Sun, Lei Xie, Eng Siong Chng, Haizhou Li, 2018. Unsupervised Domain Adaptation via Domain Adversarial Training for Speaker Recognition. Available at: http://sigport.org/3173.
Qing Wang, Wei Rao, Sining Sun, Lei Xie, Eng Siong Chng, Haizhou Li. (2018). "Unsupervised Domain Adaptation via Domain Adversarial Training for Speaker Recognition." Web.
1. Qing Wang, Wei Rao, Sining Sun, Lei Xie, Eng Siong Chng, Haizhou Li. Unsupervised Domain Adaptation via Domain Adversarial Training for Speaker Recognition [Internet]. IEEE SigPort; 2018. Available from : http://sigport.org/3173