Simplified Simultaneous Perturbation Stochastic Approximation for the Optimization of Free Decoding Parameters

  • Aleksei Romanenko
  • Alexander Zatvornitsky
  • Ivan Medennikov
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8773)


This paper deals with automatic optimization of free decoding parameters. We propose using a Simplified Simultaneous Perturbation Stochastic Approximation algorithm to optimize these parameters. This method provides a significant reduction in computational and labor costs. We also demonstrate that the proposed method successfully copes with the optimization of parameters for a specific target real-time factor, for all the databases we tested.


Simplified Simultaneous Perturbation Stochastic Approximation SPSA decoding parameter real-time factor RTF speech recognition 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Stein, D., Schwenninger, J., Stadtschnitzer, M.: Simultaneous perturbation stochastic approximation for automatic speech recognition. In: Proc. of the INTERSPEECH, Lyon, France, August 25-29, pp. 622–626 (2013)Google Scholar
  2. 2.
    El Hannani, A., Hain, T.: Automatic optimization of speech decoder parameters. Signal Processing Letters 17(1), 95–98 (2010), doi:10.1109/LSP.2009.2033967CrossRefGoogle Scholar
  3. 3.
    Mak, B., Ko, T.: Automatic estimation of decoding parameters using large-margin iterative linear programming. In: Proc. of the INTERSPEECH, Brighton, United Kingdom, September 6-10, pp. 1219–1222 (2009)Google Scholar
  4. 4.
    Kacur, J., Korosi, J.: An accuracy optimization of a dialog ASR system utilizing evolutional strategies. In: Proc. of the ISPA, Istanbul, Turkey, September 27-29, pp. 180–184 (2007)Google Scholar
  5. 5.
    Spall, J.C.: Multivariate stochastic approximation using a simultaneous perturbation gradient approximation. IEEE Transactions on Automatic Control 37(3), 332–341 (1992), doi:10.1109/9.119632MathSciNetCrossRefzbMATHGoogle Scholar
  6. 6.
    Spall, J.C.: Implementation of the simultaneous perturbation algorithm for stochastic optimization. IEEE Transactions on Aerospace and Electronic Systems 34(3), 817–823 (1998), doi:10.1109/7.705889CrossRefGoogle Scholar
  7. 7.
    Kiefer, J., Wolfowitz, J.: Stochastic Estimation of the Maximum of a Regression Function. Ann. Math. Stat. 23(3), 462–466 (1952)MathSciNetCrossRefzbMATHGoogle Scholar
  8. 8.
    Yue, X.: Improved Simultaneous Perturbation Stochastic Approximation and Its Application in Reinforcement Learning. In: Proc. of the International Conference on Computer Science and Software Engineering, Wuhan, Hubei, December 12-14, vol. 1, pp. 329–332 (2008)Google Scholar
  9. 9.
    Korenevsky, M., Bulusheva, A., Levin, K.: Unknown Words Modeling in Training and Using Language Models for Russian LVCSR System. In: Proc. of the SPECOM, Kazan, Russia, September 27-30, pp. 144–150 (2011)Google Scholar
  10. 10.
    Yurkov, P., Korenevsky, M., Levin, K.: An Improvement of robustness to speech loudness change for an ASR system based on LC-RC features. In: Proc. of the SPECOM, Kazan, Russia, September 27-30, pp. 62–66 (2011)Google Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  • Aleksei Romanenko
    • 1
  • Alexander Zatvornitsky
    • 2
  • Ivan Medennikov
    • 1
    • 3
  1. 1.ITMO UniversitySaint-PetersburgRussia
  2. 2.Speech Technology CenterSaint-PetersburgRussia
  3. 3.Saint-Petersburg State UniversitySaint-PetersburgRussia

Personalised recommendations