ISCA Archive Interspeech 2021
ISCA Archive Interspeech 2021

Correcting Automated and Manual Speech Transcription Errors Using Warped Language Models

Mahdi Namazifar, John Malik, Li Erran Li, Gokhan Tur, Dilek Hakkani Tür

Masked language models have revolutionized natural language processing systems in the past few years. A recently introduced generalization of masked language models called warped language models are trained to be more robust to the types of errors that appear in automatic or manual transcriptions of spoken language by exposing the language model to the same types of errors during the training of language models. In this work we propose a novel approach that takes advantage of the robustness of warped language models to transcription noise for correcting transcriptions of spoken language. We show that our proposed approach is able to achieve up to 10% reduction in word error rates of both automatic and manual transcriptions of spoken language.


doi: 10.21437/Interspeech.2021-591

Cite as: Namazifar, M., Malik, J., Li, L.E., Tur, G., Tür, D.H. (2021) Correcting Automated and Manual Speech Transcription Errors Using Warped Language Models. Proc. Interspeech 2021, 2037-2041, doi: 10.21437/Interspeech.2021-591

@inproceedings{namazifar21_interspeech,
  author={Mahdi Namazifar and John Malik and Li Erran Li and Gokhan Tur and Dilek Hakkani Tür},
  title={{Correcting Automated and Manual Speech Transcription Errors Using Warped Language Models}},
  year=2021,
  booktitle={Proc. Interspeech 2021},
  pages={2037--2041},
  doi={10.21437/Interspeech.2021-591}
}