ISCA Archive Interspeech 2008
ISCA Archive Interspeech 2008

Evaluating spoken language model based on filler prediction model in speech recognition

Kengo Ohta, Masatoshi Tsuchiya, Seiichi Nakagawa

We propose a method that uses a filler prediction model for building a language model that includes fillers from a corpus without fillers. In our method, a filler prediction model is trained from a corpus that does not cover domain-relevant topics. It recovers fillers in inexact transcribed corpora in the target domain, and then a language model that includes fillers is built from the corpora. The results of an evaluation of the Japanese National Diet Record showed that a model using our method achieves higher recognition performance than conventional ones.


doi: 10.21437/Interspeech.2008-256

Cite as: Ohta, K., Tsuchiya, M., Nakagawa, S. (2008) Evaluating spoken language model based on filler prediction model in speech recognition. Proc. Interspeech 2008, 1558-1561, doi: 10.21437/Interspeech.2008-256

@inproceedings{ohta08_interspeech,
  author={Kengo Ohta and Masatoshi Tsuchiya and Seiichi Nakagawa},
  title={{Evaluating spoken language model based on filler prediction model in speech recognition}},
  year=2008,
  booktitle={Proc. Interspeech 2008},
  pages={1558--1561},
  doi={10.21437/Interspeech.2008-256}
}