Confidence Estimation and Deletion Prediction Using Bidirectional Recurrent Neural Networks
2018 IEEE Spoken Language Technology Workshop, SLT 2018 - Proceedings
MetadataShow full item record
Ragni, A., Li, Q., Gales, M., & Wang, Y. (2019). Confidence Estimation and Deletion Prediction Using Bidirectional Recurrent Neural Networks. 2018 IEEE Spoken Language Technology Workshop, SLT 2018 - Proceedings, 204-211. https://doi.org/10.1109/SLT.2018.8639678
The standard approach to assess reliability of automatic speech transcriptions is through the use of confidence scores. If accurate, these scores provide a flexible mechanism to flag transcription errors for upstream and downstream applications. One challenging type of errors that recognisers make are deletions. These errors are not accounted for by the standard confidence estimation schemes and are hard to rectify in the upstream and downstream processing. High deletion rates are prominent in limited resource and highly mismatched training/testing conditions studied under IARPA Babel and Material programs. This paper looks at the use of bidirectional recurrent neural networks to yield confidence estimates in predicted as well as deleted words. A simple weighting scheme is examined for combination. To assess usefulness of this approach, the combined confidence score is examined for untranscribed data selection that favours transcriptions with lower deletion errors. Experiments are conducted using IARPA Babel/Material program languages.
ALTA Institute, Cambridge University; The Office of the Director of National Intelligence (ODNI), Intelligence Advanced Research Projects Activity (IARPA) via Air Force Research Laboratory (AFRL)
External DOI: https://doi.org/10.1109/SLT.2018.8639678
This record's URL: https://www.repository.cam.ac.uk/handle/1810/287923