Automated structure discovery and parameter tuning of neural network language model based on evolution strategy

Tomohiro Tanaka, Takafumi Moriya, Takahiro Shinozaki, Shinji Watanabe, Takaaki Hori, Kevin Duh

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Citations (Scopus)

Abstract

Long short-term memory (LSTM) recurrent neural network based language models are known to improve speech recognition performance. However, significant effort is required to optimize network structures and training configurations. In this study, we automate the development process using evolutionary algorithms. In particular, we apply the covariance matrix adaptation-evolution strategy (CMA-ES), which has demonstrated robustness in other black box hyper-parameter optimization problems. By flexibly allowing optimization of various meta-parameters including layer wise unit types, our method automatically finds a configuration that gives improved recognition performance. Further, by using a Pareto based multi-objective CMA-ES, both WER and computational time were reduced jointly: after 10 generations, relative WER and computational time reductions for decoding were 4.1% and 22.7% respectively, compared to an initial baseline system whose WER was 8.7%.

Original languageEnglish
Title of host publication2016 IEEE Workshop on Spoken Language Technology, SLT 2016 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages665-671
Number of pages7
ISBN (Electronic)9781509049035
DOIs
Publication statusPublished - 2017 Feb 7
Externally publishedYes
Event2016 IEEE Workshop on Spoken Language Technology, SLT 2016 - San Diego, United States
Duration: 2016 Dec 132016 Dec 16

Publication series

Name2016 IEEE Workshop on Spoken Language Technology, SLT 2016 - Proceedings

Other

Other2016 IEEE Workshop on Spoken Language Technology, SLT 2016
CountryUnited States
CitySan Diego
Period16/12/1316/12/16

Keywords

  • Evolution strategy
  • Language model
  • Large vocabulary speech recognition
  • Long short-term memory
  • Multi-objective optimization

ASJC Scopus subject areas

  • Human-Computer Interaction
  • Artificial Intelligence
  • Language and Linguistics
  • Computer Vision and Pattern Recognition
  • Computer Science Applications

Fingerprint Dive into the research topics of 'Automated structure discovery and parameter tuning of neural network language model based on evolution strategy'. Together they form a unique fingerprint.

Cite this