ISCA Archive Interspeech 2013
ISCA Archive Interspeech 2013

Simultaneous perturbation stochastic approximation for automatic speech recognition

Daniel Stein, Jochen Schwenninger, Michael Stadtschnitzer

While both the acoustic model and the language model in automatic speech recognition are typically well-trained on the target domain, the free parameters of the decoder itself are often set manually. In this paper, we investigate in how far a stochastic approximation algorithm can be employed to automatically determine the best parameters, especially if additional time-constraints are given on unknown machine architectures. We offer our findings on the German Difficult Speech Corpus, and present significant improvements over both the spontaneous and planned clean speech task.


doi: 10.21437/Interspeech.2013-181

Cite as: Stein, D., Schwenninger, J., Stadtschnitzer, M. (2013) Simultaneous perturbation stochastic approximation for automatic speech recognition. Proc. Interspeech 2013, 622-626, doi: 10.21437/Interspeech.2013-181

@inproceedings{stein13_interspeech,
  author={Daniel Stein and Jochen Schwenninger and Michael Stadtschnitzer},
  title={{Simultaneous perturbation stochastic approximation for automatic speech recognition}},
  year=2013,
  booktitle={Proc. Interspeech 2013},
  pages={622--626},
  doi={10.21437/Interspeech.2013-181}
}