PHONE RECOGNITION WITH DEEP SPARSE RECTIFIER NEURAL NETWORKS

被引：0
|
作者：
Toth, Laszlo [1 ]
机构：
[1] Hungarian Acad Sci, MTA SZTE Res Grp Artificial Intelligence, H-1051 Budapest, Hungary
来源：
2013 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP) | 2013年
关键词：
Deep neural networks; sparse rectifier neural networks; phone recognition;
D O I：
暂无
中图分类号：
O42 [声学];
学科分类号：
070206 ; 082403 ;
摘要：
Rectifier neurons differ from standard ones only in that the sigmoid activation function is replaced by the rectifier function, max(0, x). This modification requires only minimal changes to any existing neural net implementation, but makes it more effective. In particular, we show that a deep architecture of rectifier neurons can attain the same recognition accuracy as deep neural networks, but without the need for pre-training. With 4-5 hidden layers of rectifier neurons we report 20.8% and 19.8% phone error rates on TIMIT (with CI and CD units, respectively), which are competitive with the best results on this database.
引用
收藏
页码：6985 / 6989
页数：5
相关论文
共 50 条
  • [1] Convolutional Deep Rectifier Neural Nets for Phone Recognition
    Toth, Laszlo
    [J]. 14TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2013), VOLS 1-5, 2013, : 1721 - 1725
  • [2] DEEP SPARSE RECTIFIER NEURAL NETWORKS FOR SPEECH DENOISING
    Xu, Lie
    Choy, Chiu-Sing
    Li, Yi-Wen
    [J]. 2016 IEEE INTERNATIONAL WORKSHOP ON ACOUSTIC SIGNAL ENHANCEMENT (IWAENC), 2016,
  • [3] INVESTIGATING SPARSE DEEP NEURAL NETWORKS FOR SPEECH RECOGNITION
    Pironkov, Gueorgui
    Dupont, Stephane
    Dutoit, Thierry
    [J]. 2015 IEEE WORKSHOP ON AUTOMATIC SPEECH RECOGNITION AND UNDERSTANDING (ASRU), 2015, : 124 - 129
  • [4] Deep Neural Networks with Linearly Augmented Rectifier Layers for Speech Recognition
    Toth, Laszlo
    [J]. 2018 IEEE 16TH WORLD SYMPOSIUM ON APPLIED MACHINE INTELLIGENCE AND INFORMATICS (SAMI 2018): DEDICATED TO THE MEMORY OF PIONEER OF ROBOTICS ANTAL (TONY) K. BEJCZY, 2018, : 189 - 193
  • [5] A Sequence Training Method for Deep Rectifier Neural Networks in Speech Recognition
    Grosz, Tamas
    Gosztolya, Gabor
    Toth, Laszlo
    [J]. SPEECH AND COMPUTER, 2014, 8773 : 81 - 88
  • [6] Weather recognition of street scene based on sparse deep neural networks
    Liu, Wei
    Yang, Yue
    Wei, Longsheng
    [J]. Journal of Advanced Computational Intelligence and Intelligent Informatics, 2017, 21 (03) : 403 - 408
  • [7] Hyperspectral face recognition based on sparse spectral attention deep neural networks
    Xie, Zhihua
    Li, Yi
    Niu, Jieyi
    Shi, Ling
    Wang, Zhipeng
    Lu, Guoyu
    [J]. OPTICS EXPRESS, 2020, 28 (24) : 36286 - 36303
  • [8] Convolutional Deep Maxout Networks for Phone Recognition
    Toth, Laszlo
    [J]. 15TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2014), VOLS 1-4, 2014, : 1078 - 1082
  • [9] Spoken Letter Recognition using Deep Convolutional Neural Networks on Sparse and Dissimilar Data
    Kalischewski, Kathrin
    Wagner, Daniel
    Velten, Joerg
    Kummert, Anton
    [J]. 2019 IEEE INTERNATIONAL SYMPOSIUM ON CIRCUITS AND SYSTEMS (ISCAS), 2019,
  • [10] Improved automatic speech recognition system using sparse decomposition by basis pursuit with deep rectifier neural networks and compressed sensing recomposition of speech signals
    Gavrilescu, Mihai
    [J]. 2014 10TH INTERNATIONAL CONFERENCE ON COMMUNICATIONS (COMM), 2014,
12345