Adversarially Learned Total Variability Embedding for Speaker Recognition with Random Digit Strings
Over the recent years, various research has been conducted to investigate methods for verifying users with a short randomized pass-phrase due to the increasing demand for voice-based authentication systems. In this paper, we propose a novel technique for extracting an i-vector-like feature based on...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2019-10-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/19/21/4709 |
| _version_ | 1817990869925494784 |
|---|---|
| author | Woo Hyun Kang Nam Soo Kim |
| author_facet | Woo Hyun Kang Nam Soo Kim |
| author_sort | Woo Hyun Kang |
| collection | DOAJ |
| description | Over the recent years, various research has been conducted to investigate methods for verifying users with a short randomized pass-phrase due to the increasing demand for voice-based authentication systems. In this paper, we propose a novel technique for extracting an i-vector-like feature based on an adversarially learned inference (ALI) model which summarizes the variability within the Gaussian mixture model (GMM) distribution through a nonlinear process. Analogous to the previously proposed variational autoencoder (VAE)-based feature extractor, the proposed ALI-based model is trained to generate the GMM supervector according to the maximum likelihood criterion given the Baum−Welch statistics of the input utterance. However, to prevent the potential loss of information caused by the Kullback−Leibler divergence (KL divergence) regularization adopted in the VAE-based model training, the newly proposed ALI-based feature extractor exploits a joint discriminator to ensure that the generated latent variable and the GMM supervector are more realistic. The proposed framework is compared with the conventional i-vector and VAE-based methods using the TIDIGITS dataset. Experimental results show that the proposed method can represent the uncertainty caused by the short duration better than the VAE-based method. Furthermore, the proposed approach has shown great performance when applied in association with the standard i-vector framework. |
| first_indexed | 2024-04-14T01:04:46Z |
| format | Article |
| id | doaj.art-c3f4dfedea4945ec9df2adf204542370 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-04-14T01:04:46Z |
| publishDate | 2019-10-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-c3f4dfedea4945ec9df2adf2045423702022-12-22T02:21:17ZengMDPI AGSensors1424-82202019-10-011921470910.3390/s19214709s19214709Adversarially Learned Total Variability Embedding for Speaker Recognition with Random Digit StringsWoo Hyun Kang0Nam Soo Kim1Department of Electrical and Computer Engineering and the Institute of New Media and Communications, Seoul National University, Seoul 08826, KoreaDepartment of Electrical and Computer Engineering and the Institute of New Media and Communications, Seoul National University, Seoul 08826, KoreaOver the recent years, various research has been conducted to investigate methods for verifying users with a short randomized pass-phrase due to the increasing demand for voice-based authentication systems. In this paper, we propose a novel technique for extracting an i-vector-like feature based on an adversarially learned inference (ALI) model which summarizes the variability within the Gaussian mixture model (GMM) distribution through a nonlinear process. Analogous to the previously proposed variational autoencoder (VAE)-based feature extractor, the proposed ALI-based model is trained to generate the GMM supervector according to the maximum likelihood criterion given the Baum−Welch statistics of the input utterance. However, to prevent the potential loss of information caused by the Kullback−Leibler divergence (KL divergence) regularization adopted in the VAE-based model training, the newly proposed ALI-based feature extractor exploits a joint discriminator to ensure that the generated latent variable and the GMM supervector are more realistic. The proposed framework is compared with the conventional i-vector and VAE-based methods using the TIDIGITS dataset. Experimental results show that the proposed method can represent the uncertainty caused by the short duration better than the VAE-based method. Furthermore, the proposed approach has shown great performance when applied in association with the standard i-vector framework.https://www.mdpi.com/1424-8220/19/21/4709speech embeddingdeep learningspeaker recognitionunsupervised representation learning |
| spellingShingle | Woo Hyun Kang Nam Soo Kim Adversarially Learned Total Variability Embedding for Speaker Recognition with Random Digit Strings Sensors speech embedding deep learning speaker recognition unsupervised representation learning |
| title | Adversarially Learned Total Variability Embedding for Speaker Recognition with Random Digit Strings |
| title_full | Adversarially Learned Total Variability Embedding for Speaker Recognition with Random Digit Strings |
| title_fullStr | Adversarially Learned Total Variability Embedding for Speaker Recognition with Random Digit Strings |
| title_full_unstemmed | Adversarially Learned Total Variability Embedding for Speaker Recognition with Random Digit Strings |
| title_short | Adversarially Learned Total Variability Embedding for Speaker Recognition with Random Digit Strings |
| title_sort | adversarially learned total variability embedding for speaker recognition with random digit strings |
| topic | speech embedding deep learning speaker recognition unsupervised representation learning |
| url | https://www.mdpi.com/1424-8220/19/21/4709 |
| work_keys_str_mv | AT woohyunkang adversariallylearnedtotalvariabilityembeddingforspeakerrecognitionwithrandomdigitstrings AT namsookim adversariallylearnedtotalvariabilityembeddingforspeakerrecognitionwithrandomdigitstrings |