Neural Architecture Search For Keyword Spotting

被引:16
|
作者
Mo, Tong [1 ]
Yu, Yakun [1 ]
Salameh, Mohammad [2 ]
Niu, Di [1 ]
Jui, Shangling [2 ]
机构
[1] Univ Alberta, Edmonton, AB, Canada
[2] Huawei Technol, Shenzhen, Peoples R China
来源
INTERSPEECH 2020 | 2020年
关键词
Keyword Spotting; Neural Architecture Search;
D O I
10.21437/Interspeech.2020-3132
中图分类号
R36 [病理学]; R76 [耳鼻咽喉科学];
学科分类号
100104 ; 100213 ;
摘要
Deep neural networks have recently become a popular solution to keyword spotting systems, which enable the control of smart devices via voice. In this paper, we apply neural architecture search to search for convolutional neural network models that can help boost the performance of keyword spotting based on features extracted from acoustic signals while maintaining an acceptable memory footprint. Specifically, we use differentiable architecture search techniques to search for operators and their connections in a predefined cell search space. The found cells are then scaled up in both depth and width to achieve competitive performance. We evaluated the proposed method on Google's Speech Commands Dataset and achieved a state-of-the-art accuracy of over 97% on the setting of 12-class utterance classification commonly reported in the literature.
引用
收藏
页码:1982 / 1986
页数:5
相关论文
共 50 条
  • [1] STOCHASTIC ADAPTIVE NEURAL ARCHITECTURE SEARCH FOR KEYWORD SPOTTING
    Veniat, Tom
    Schwander, Olivier
    Denoyer, Ludovic
    2019 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2019, : 2842 - 2846
  • [2] Target-Aware Neural Architecture Search and Deployment for Keyword Spotting
    Busia, Paola
    Deriu, Gianfranco
    Rinelli, Luca
    Chesta, Cristina
    Raffo, Luigi
    Meloni, Paolo
    IEEE ACCESS, 2022, 10 : 40687 - 40700
  • [3] AUTOKWS: KEYWORD SPOTTING WITH DIFFERENTIABLE ARCHITECTURE SEARCH
    Zhang, Bo
    Li, Wenfeng
    Li, Qingyuan
    Zhuang, Weiji
    Chu, Xiangxiang
    Wang, Yujun
    2021 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP 2021), 2021, : 2830 - 2834
  • [4] Improving Keyword Spotting and Language Identification via Neural Architecture Search at Scale
    Mazzawi, Hanna
    Gonzalvo, Xavi
    Kracun, Aleks
    Sridhar, Prashant
    Subrahmanya, Niranjan
    Moreno, Ignacio Lopez
    Park, Hyun Jin
    Violette, Patrick
    INTERSPEECH 2019, 2019, : 1278 - 1282
  • [5] END-TO-END KEYWORD SPOTTING USING NEURAL ARCHITECTURE SEARCH AND QUANTIZATION
    Peter, David
    Roth, Wolfgang
    Pernkopf, Franz
    2022 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2022, : 3423 - 3427
  • [6] Resource-efficient DNNs for Keyword Spotting using Neural Architecture Search and Quantization
    Peter, David
    Roth, Wolfgang
    Pernkopf, Franz
    2020 25TH INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION (ICPR), 2021, : 9273 - 9279
  • [7] Channel Shuffle Neural Architecture Search for Key Word Spotting
    Lee, Bokyeung
    Kim, Donghyeon
    Kim, Gwantae
    Ko, Hanseok
    IEEE SIGNAL PROCESSING LETTERS, 2023, 30 : 443 - 447
  • [8] Keyword spotting based on recurrent neural network
    Zhou, JL
    Liu, J
    Song, YT
    Yu, TC
    ICSP '98: 1998 FOURTH INTERNATIONAL CONFERENCE ON SIGNAL PROCESSING, PROCEEDINGS, VOLS I AND II, 1998, : 710 - 713
  • [9] Keyword spotting based on recurrent neural network
    Chinese Acad of Science, Beijing, China
    Int Conf Signal Process Proc, (710-713):
  • [10] Audio Search Based on Keyword Spotting in Arabic Language
    Awaid, Mostafa
    Fawzi, Sahar A.
    Kandil, Ahmed H.
    INTERNATIONAL JOURNAL OF ADVANCED COMPUTER SCIENCE AND APPLICATIONS, 2014, 5 (02) : 128 - 133