DNN-Based Voice Activity Detection with Multi-Task Learning

Tae Gyoon KANG; Nam Soo KIM

doi:10.1587/transinf.2015EDL8168

DNN-Based Voice Activity Detection with Multi-Task Learning

Tae Gyoon KANG, Nam Soo KIM

Full Text Views

0

Share
Cite this

Summary :

Recently, notable improvements in voice activity detection (VAD) problem have been achieved by adopting several machine learning techniques. Among them, the deep neural network (DNN) which learns the mapping between the noisy speech features and the corresponding voice activity status with its deep hidden structure has been one of the most popular techniques. In this letter, we propose a novel approach which enhances the robustness of DNN in mismatched noise conditions with multi-task learning (MTL) framework. In the proposed algorithm, a feature enhancement task for speech features is jointly trained with the conventional VAD task. The experimental results show that the DNN with the proposed framework outperforms the conventional DNN-based VAD algorithm.

Publication: IEICE TRANSACTIONS on Information Vol.E99-D No.2 pp.550-553

Publication Date: 2016/02/01

Publicized: 2015/10/26

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2015EDL8168

Type of Manuscript: LETTER

Category: Speech and Hearing

Authors

Tae Gyoon KANG
Seoul National University
Nam Soo KIM
Seoul National University

Keyword

deep neural network, voice activity detection, multi-task learning

Cite this

Copy

Tae Gyoon KANG, Nam Soo KIM, "DNN-Based Voice Activity Detection with Multi-Task Learning" in IEICE TRANSACTIONS on Information, vol. E99-D, no. 2, pp. 550-553, February 2016, doi: 10.1587/transinf.2015EDL8168.
Abstract: Recently, notable improvements in voice activity detection (VAD) problem have been achieved by adopting several machine learning techniques. Among them, the deep neural network (DNN) which learns the mapping between the noisy speech features and the corresponding voice activity status with its deep hidden structure has been one of the most popular techniques. In this letter, we propose a novel approach which enhances the robustness of DNN in mismatched noise conditions with multi-task learning (MTL) framework. In the proposed algorithm, a feature enhancement task for speech features is jointly trained with the conventional VAD task. The experimental results show that the DNN with the proposed framework outperforms the conventional DNN-based VAD algorithm.
URL: https://globals.ieice.org/en_transactions/information/10.1587/transinf.2015EDL8168/_p

Copy

@ARTICLE{e99-d_2_550,
author={Tae Gyoon KANG, Nam Soo KIM, },
journal={IEICE TRANSACTIONS on Information},
title={DNN-Based Voice Activity Detection with Multi-Task Learning},
year={2016},
volume={E99-D},
number={2},
pages={550-553},
abstract={Recently, notable improvements in voice activity detection (VAD) problem have been achieved by adopting several machine learning techniques. Among them, the deep neural network (DNN) which learns the mapping between the noisy speech features and the corresponding voice activity status with its deep hidden structure has been one of the most popular techniques. In this letter, we propose a novel approach which enhances the robustness of DNN in mismatched noise conditions with multi-task learning (MTL) framework. In the proposed algorithm, a feature enhancement task for speech features is jointly trained with the conventional VAD task. The experimental results show that the DNN with the proposed framework outperforms the conventional DNN-based VAD algorithm.},
keywords={},
doi={10.1587/transinf.2015EDL8168},
ISSN={1745-1361},
month={February},}

Copy

TY - JOUR
TI - DNN-Based Voice Activity Detection with Multi-Task Learning
T2 - IEICE TRANSACTIONS on Information
SP - 550
EP - 553
AU - Tae Gyoon KANG
AU - Nam Soo KIM
PY - 2016
DO - 10.1587/transinf.2015EDL8168
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E99-D
IS - 2
JA - IEICE TRANSACTIONS on Information
Y1 - February 2016
AB - Recently, notable improvements in voice activity detection (VAD) problem have been achieved by adopting several machine learning techniques. Among them, the deep neural network (DNN) which learns the mapping between the noisy speech features and the corresponding voice activity status with its deep hidden structure has been one of the most popular techniques. In this letter, we propose a novel approach which enhances the robustness of DNN in mismatched noise conditions with multi-task learning (MTL) framework. In the proposed algorithm, a feature enhancement task for speech features is jointly trained with the conventional VAD task. The experimental results show that the DNN with the proposed framework outperforms the conventional DNN-based VAD algorithm.
ER -