Art

J-GLOBAL ID：202002244722287209   Reference number：20A2051023

Evolutionary Development of Transformer Speech Recognition System.

Transformer音声認識システムの進化的最適化

Author (2)： HINO KENTO (東京工大) ,  SHINOZAKI TAKAHIRO (東京工大)
Material： 日本音響学会研究発表会講演論文集(CD-ROM)  (日本音響学会講演論文集(CD-ROM))
Volume： 2020  Issue： 秋季  Page： ROMBUNNO.2-P1-6  Publication year： Aug. 26, 2020 
JST Material Number： G0381C  ISSN： 1880-7658  Document type： Proceedings
Article type： 短報  Country of issue： Japan (JPN)  Language： JAPANESE (JA)

Thesaurus term： optimization ,  *speech recognition ,  *structural optimization ,  neural network (biomodel) ,  optimization method ,  *speech-recognition system ,  evolution strategy ,  *Evolutionary optimization ,  *modification
Semi thesaurus term： *Transformer音声認識システム

JST classification (1)： Pattern recognition  (JE07000S)

Reference (5)：

• N. Hansen, S. D. Müller and P. Koumoutsakos, “Reducing the time complexity of the derandomized evolution strategy with covariance matrix adaptation (CMA-ES),”Proc. Evolutionary Computation, 2003.
• Takahiro Shinozaki, Shinji Watanabe, ”Structure discovery of deep neural network based on evolutionary algorithms,” Proc. ICASSP, 2015.4
• Kent Hino, Yusuke Kimura, Yue Dong and Takahiro Shinozaki, ”Dual Inheritance Evolution Strategy for Deep Neural Network Optimization,” Proc. IEEE Congress on Evolution Computation (CEC) 2020, accepted.
• Adriana Romero, et al., ”FitNets: Hints for Thin Deep Nets”, 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings (2015).
• Shinji Watanabe, et al., ”ESPnet: End-to-End Speech Processing Toolkit”, Proc. Interspeech, 2018

Terms in the title (3)： Transformer ,  音声認識システム ,  進化的最適化