Improving Semi-supervised Deep Neural Network. for Keyword Search in Low Resource Languages

被引：0
|
作者：
Hsiao, Roger [1 ]
Ng, Tim [1 ]
Le Zhang [1 ]
Ranjan, Shivesh [1 ]
机构：
[1] Raytheon BBN Technol, 10 Moulton St, Cambridge, MA 02138 USA
来源：
15TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2014), VOLS 1-4 | 2014年
关键词：
semi-supervised training; deep neural network; keyword search;
D O I：
暂无
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
In this work, we investigate how to improve semi-supervised DNN for low resource languages where the initial systems may have high error rate. We propose using semi-supervised MLP features for DNN training, and we also explore using confidence to improve semi-supervised cross entropy and sequence training. The work conducted in this paper was evaluated under the IARPA Babel program for the keyword spotting tasks. We focus on the limited condition where there are around 10 hours of supervised data for training.
引用
收藏
页码：1088 / 1091
页数：4
相关论文
共 36 条
  • [1] DISCRIMINATIVE SEMI-SUPERVISED TRAINING FOR KEYWORD SEARCH IN LOW RESOURCE LANGUAGES
    Hsiao, Roger
    Ng, Tim
    Grezl, Frantisek
    Karakos, Damianos
    Tsakalidis, Stavros
    Long Nguyen
    Schwartz, Richard
    2013 IEEE WORKSHOP ON AUTOMATIC SPEECH RECOGNITION AND UNDERSTANDING (ASRU), 2013, : 440 - 445
  • [2] DEEP NEURAL NETWORK FEATURES AND SEMI-SUPERVISED TRAINING FOR LOW RESOURCE SPEECH RECOGNITION
    Thomas, Samuel
    Seltzer, Michael L.
    Church, Kenneth
    Hermansky, Hynek
    2013 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2013, : 6704 - 6708
  • [3] Improving Speech Recognition and Keyword Search for Low Resource Languages Using Web Data
    Mendels, Gideon
    Cooper, Erica
    Soto, Victor
    Hirschberg, Julia
    Gales, Mark
    Knill, Kate
    Ragni, Anton
    Wang, Haipeng
    16TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2015), VOLS 1-5, 2015, : 829 - 833
  • [4] Multi-softmax Deep Neural Network for Semi-supervised Training
    Su, Hang
    Xu, Haihua
    16TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2015), VOLS 1-5, 2015, : 3239 - 3243
  • [5] SEMI-SUPERVISED TERM-WEIGHTED VALUE RESCORING FOR KEYWORD SEARCH
    Audhkhasi, Kartik
    Sethy, Abhinav
    Ramabhadran, Bhuvana
    Narayanan, Shrikanth S.
    2014 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2014,
  • [6] SEMI-SUPERVISED TRAINING OF DEEP NEURAL NETWORKS
    Vesely, Karel
    Hannemann, Mirko
    Burget, Lukas
    2013 IEEE WORKSHOP ON AUTOMATIC SPEECH RECOGNITION AND UNDERSTANDING (ASRU), 2013, : 267 - 272
  • [7] Labeling Malicious Communication Samples Based on Semi-Supervised Deep Neural Network
    Shao, Guolin
    Chen, Xingshu
    Zeng, Xuemei
    Wang, Lina
    CHINA COMMUNICATIONS, 2019, 16 (11) : 183 - 200
  • [8] Labeling Malicious Communication Samples Based on Semi-Supervised Deep Neural Network
    Guolin Shao
    Xingshu Chen
    Xuemei Zeng
    Lina Wang
    中国通信, 2019, 16 (11) : 183 - 200
  • [9] Limited Data Spectrum Sensing Based on Semi-Supervised Deep Neural Network
    Zhang, Yupei
    Zhao, Zhijin
    IEEE ACCESS, 2021, 9 : 166423 - 166435
  • [10] Semi-supervised Maximum Mutual Information Training of Deep Neural Network Acoustic Models
    Manohar, Vimal
    Povey, Daniel
    Khudanpur, Sanjeev
    16TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2015), VOLS 1-5, 2015, : 2630 - 2634
1234